CHAPTER I RIASSUNTO 1 ABSTRACT 9 INTRODUCTION 10 Motility and Inflammation 10 AIMS 15

DEVELOPMENT OF A CELL MOTILITY CHARACTERIZATION SYSTEM FOR INDUSTRIAL BIOTECHNOLOGICAL APPLICATIONS

Valeria Rachela Villella

DEVELOPMENT OF A CELL MOTILITY CHARACTERIZATION SYSTEM FOR INDUSTRIAL BIOTECHNOLOGICAL APPLICATIONS

Valeria Rachela Villella Dottoranda: Valeria Rachela Villella

Relatore: Prof. Stefano Guido

Coordinatore: Prof. Giovanni Sannia

La grandezza di una persona si evince nella perseveranza e nella fermezza che investe nel perseguire e realizzare i suoi grandi sogni.

A chi ha creduto in me

INDICE

CHAPTER I RIASSUNTO 1 ABSTRACT 9 INTRODUCTION 10 Motility and Inflammation 10 AIMS 15

CHAPTER II RESULTS 16 1.The inflammatory environment of Cystic Fibrosis 16 1.1.The up-regulation of TG2 in human CFTR-defective cells 16 1.2.SUMOylation of Tissue Transglutaminase 18 1.3.Imapct of TG2 persistence in CF epithelia 23 1.4.Aggresome: a feature of CF epithelia 27 1.5.Aggresome and autophagy 28 1.6.In vivo model 36 2. Chemotaxis: Motility model 41 2.1.The 3D matrix collagen gel 41 2.2.TIME-LAPSE system 44 2.3.Calibration of system 46 2.4.Neutrophils in collagen gel: 3D model 48 2.5.T84 motility: 2D model 50

CHAPTER III CONCLUSIONS 55

CHAPTER IV MATERIALS AND METHODS 58 REFERENCES 67 APPENDIX 77

CHAPTER I

RIASSUNTO

Premesse scientifiche

I fenomeni di interazione e migrazione cellulare sono rilevanti in diversi processi fisiopatologici dallo sviluppo embrionale all’infiammazione. Le tecnologie finora disponibili per lo studio di tali fenomeni in vitro sono in larga parte non adatte per applicazioni biotecnologiche industriali. Infatti, l’analisi delle interazioni cellulari viene spesso effettuata con saggi di biologia cellulare quali osservazioni in immunofluorescenza di campioni fissati, che sono limitate a condizioni statiche e si prestano con difficoltà a fornire misure quantitative di eventi dinamici, come la motilità cellulare, su larga scala. Scopo del progetto di tesi è lo sviluppo di un sistema innovativo per l’analisi della motilità e delle interazioni cellulari in 2D e 3D che, superando i limiti già accennati dei saggi finora disponibili, si presti a utilizzo biotecnologico industriale per future applicazioni in screening di sostanze a potenziale valenza terapeutica. In particolare, è stato sviluppato un sistema che consente l’analisi quantitativa del fenomeno di chemiotassi, in cui un gradiente di chemochine, attraverso recettori di superficie, innesca cascate di segnali intracellulari che culminano nel movimento di cellule infiammatorie in direzione del chemoattrattante (Hynes R.,Cell 2002). Si tratta di un fenomeno in cui una costante e dinamica interazione fra il microambiente e la superficie cellulare induce modificazioni dello stesso fenotipo della cellula che risponde a questi stimoli con movimenti finalistici (Ridley A, et al. Science 2003). Cio' avviene, ad esempio, per il movimento dei neutrofili in risposta all’interlechina 8 (IL-8). Il lavoro di tesi è stato organizzato nel modo seguente. In un prima fase l’attività di ricerca è stata indirizzata all’individuazione di meccanismi molecolari che regolano la risposta infiammatoria alla base della produzione di fattori chemoattraenti, come IL-8, utilizzando in particolare come modello la fibrosi cistica. Questa prima parte del lavoro di tesi è stata condotta con tecniche classiche di biologia cellulare e molecolare ed è stata finalizzata alla comprensione del meccanismo molecolare alla base della produzione di IL-8. E’ stato studiato l’effetto di un ambiente infiammatorio su fattori come la transglutaminasi (TG2) che a sua volta regola la modulazione di IL- 8. I risultati ottenuti in questa prima parte del lavoro sono stati utilizzati per la messa a punto di una cella di chemiotassi in cui neutrofili migrano in un gel di collagene tridimensionale sotto l’effetto di un gradiente di IL-8. E’ stato inoltre messo a punto un sistema di valenza trasversale per l’analisi quantitativa della migrazione collettiva di una linea cellulare (T84, utilizzata come modello nella patologia celiaca) durante il processo di confluenza in una piastra di coltura 2D. Si tratta di un ulteriore esempio dell'interazione dinamica fra cellule attraverso il quale cellule poste in piastre di coltura in apposito terreno raggiungono la confluenza muovendosi in uno spazio 2D. Questo fenomeno e' particolarmente studiato per le cellule epiteliali ed e' coordinato non solo da non ancora ben definiti stimoli esterni, ma soprattutto da modificazioni dinamiche della cellula, come la proliferazione e il riarrangiamento del citoschelestro che consentono alle cellule di entrare in contatto con le cellule vicine. La complessità di questi eventi e' ulteriormente provata dal fatto che le cellule stesse, interagendo con le cellule vicine vanno incontro a differenziamento come nel caso delle linee cellulari epiteliali intestinali T84 in cui la confluenza induce produzione di enzimi di superficie, quali la lattasi, propri delle cellule mature. Questo processo di migrazione

CHAPTER I collettiva viene inoltre influenzato dagli stessi fattori, studiati nella prima parte del lavoro di tesi, che regolano la produzione di IL-8. Le motivazioni per lo sviluppo del sistema biotecnologico di analisi dei fenomeni di interazione e motilità cellulare oggetto di questa tesi sono da ricercarsi in ambito patologico. Infatti, nel panorama cosi' complesso della motilità e interazione fra cellule in un determinato microambiente la ricerca dei fattori che determinano le modifiche cellulari che conducono all'interazione rappresenta un elemento di grande interesse in patologia umana in quanto il decorso di malattie croniche e' spesso influenzato da alterazioni dei meccanismi di interazione fra cellule. Un esempio e' fornito dal comportamento cellulare in patologie neoplastiche in cui il controllo della invasività delle cellule cancerose rappresenta il goal di terapie anti-tumorali. Infine lo studio del "movimento" e della sua direzionalità in patologia umana puo' non riguardare cellule in toto o sistemi di cellule in un ambiente specifico. La "motilità" e' infatti anche una caratteristica del compartimento intracellulare in cui organelli, aggregati di proteine e le stesse proteine si muovono all'interno della cellula secondo regole ben definite. Un esempio e' rappresentato dal trafficking di proteine in membrana e dal loro continuo reciclying fra la membrana e gli endosomi nonche' la progressione verso i compartimenti degradativi intracellulari. Meccanismi di "movimenti" intracellulari sono alla base di complessi meccanismi di degradazione quali l'autofagia e compoprtano anche un riarrangiamento del citoscheletro che regola la motilità "esterna" dell'intera cellula. In un panorama cosi' complesso l'identificazione dei fattori che possono disregolare il complesso sistema della motilità in senso lato e' pertanto di grande interesse e la identificazione di molecole potenzialmente capaci di modulare questi processi puo' avere enormi implicazioni terapeutiche per applicazioni biotecnologiche industriali. La scelta di modelli di malattia idonei allo studio dei fattori putativamente coinvolti nei meccanismi di motilità/interazione cellulare dovrebbe basarsi essenzialmente sulla evidenza che questi meccanismi costituiscono un elemento patogenetico della malattia e non un semplice evento bystander. La comprensione di una pathway specifica puo' aprire la strada alla identificazione di sostanze capaci di modulare questi processi biologici e rappresentare molecole eventualmente implementabili nella terapia di molte patologie umane. Come già accennato, nel corso del dottorato di ricerca sono stati utilizzati diversi modelli cellulari di due differenti patologie umane, la Fibrosi Cistica e la Celiachia che hanno come fattore comune valori elevati della Transglutaminasi tissutale (TG2), un enzima multifunzionale con un ruolo patogenetico noto in diverse patologie umane, che induce cross-linking di proteine substato o espleta altre funzioni quali una attività isomerasica o di G protein (Malorni, W., et al. 2008, Curr. Pharm. Des). La Fibrosi Cistica, la piu' comune malattia monogenica ad esito letale nella popolazione caucasica, dovuta a mutazioni del gene CFTR, e' caratterizzata, fra l'altro, da infiammazione polmonare a prevalente componente neutrofilica in risposta ad elevata secrezione di IL8 con ricorrenti infezioni batteriche (Smith, J. J., et al. Cell. 1996; Ratjen, F., and G. Doring. 2003. Lancet). Nella prima parte del lavoro di tesi, sono state utilizzate linee cellulari epiteliali per identificare la pathway mediata da TG2 che porta ad incrementati livelli di IL8 e di infiammazione. Sono stati quindi validati inibitori specifici di questa pathway ed individuato il meccanismo di controllo dell'infiammazione e della secrezione di IL8 in vitro e in vivo in 2 diversi modelli animali di malattia. La celiachia, una intolleranza permanente a determinate sequenze peptidiche della gliadina del frumento e delle prolamine di orzo e segale, e' una comune intolleranza

CHAPTER I alimentare con una prevalenza di 1:100 nella popolazione generale, e caratterizata da una disregolazione della risposta immune mucosale innata ed adattativa con componente autoimmune (Sollid LM. Nat Rev Immunol 2002). E' ampiamente dimostrato come la gliadina induca aumentata espressione e attivazione della proteina TG2 che esercita un ruolo patogenetico nella malattia. Nella prima parte del lavoro di tesi, sono state utilizzate linee cellulari epiteliali intestinali per studiare i meccanismi di internalizzazione e degradazione di peptidi della gliadina e il loro impatto sulla upregolazione della TG2. Infine sono state utilizzate queste acquisizioni per studiare gli effetti dei peptidi tossici della gliadina sulla motilità delle cellule epiteliali in coltura e sulle loro interazioni e per comprendere se l'inibizione della TG2 possa esercitare un ruolo chiave nel controllare gli effetti dei peptidi della gliadina sulla motilità delle cellule epiteliali in un sistema 2D.

Risultati ottenuti e metodologie

La Fibrosi Cistica (FC) è un nota malattia genetica caratterizzata da uno stato infiammatorio basale, la cui natura è ancora oggetto di controversie. A) Al fine di comprendere il link tra difetto della Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) e infiammazione con elevati livelli del fattore chemoattrattante IL-8, sono stati condotti esperimenti su linee epiteliali portatrici della piu' comune mutazioni del gene CFTR, la F508del-CFTR, le IB3 e su le corrispondenti linee epiteliali in cui il difetto è stato geneticamente corretto, le linee isogeniche C38 (Smith, J. J., et al. Cell. 1996). Esperimenti di Western Blot, Immunoprecipitazione, siRNA, Microscopia confocale ed analisi dell'attività enzimatica hanno permesso di evidenziare che nelle cellule epiteliali FC gli elevati livelli di TG2 e la sua conseguente attività enzimatica sono delle funzione della CFTR in quanto sia inibitori della funzione di CFTR, quali il CFTR-inhibitor-172 che il silenziamento del gene CFTR mediante siRNA, aumentano i livelli di TG2 nelle cellule C38. Abbiamo poi dimostrato che l'inibizione della TG2, sia mediante silenziamento genico che con inibitori della sua attività enzimatica (cystamina o R283), riduce drammaticamente alcuni marcatori di attivazione epiteliale, quali la fosforilazione delle MAP-chinasi ERK1/2 (p42/44), e la produzione di citochine infiammatorie, quali TNF-a, sia della chemochina IL-8. Questi dati suggeriscono che il difetto di CFTR induce infiammazione e produzione di IL-8 mediati da TG2. B) Abbiamo poi voluto comprendere come il difetto di CFTR induca persistenza di elevati livelli di TG2 nelle cellule epiteliali FC. E' noto che la difettiva funzione di CFTR induce un aumento delle Specie Reattive all'Ossigeno (ROS) (Yi, S. J., et al.. 2006. Mol. Cells). Abbiamo dimostrato che molecole scavenger di ROS, quali la N- acetyl-cysteine (NAC) e la catalasi Superossido-Dismutasi(SOD)-mimetica EUK-134 riducono infatti i livelli di TG2 con drammatica riduzione della sua attività enzimatica. La dipendenza di elevati livelli di TG2 dall' ambiante pro-ossidativo ci ha indotto a pensare che l'incremento di ROS potesse indurre modifiche posttraslazionali della proteina TG2. Fra le piu' comuni modifiche post-traduzionali sono la Sumoylazione (Meulmeester, E., and F. Melchior. 2008. Nature), e l’ ubiquitinazione nota per intervenire in processi di degradazione via proteasoma. Entrambe competono per gli stessi residui di Lys e quindi sono mutualmente esclusive e richiedono enzimi E3- ligasi specifici.

CHAPTER I

I risultati ottenuti in seguito a IP di TG2 in cellule IB3 e C38 hanno evidenziato elevati livelli di Sumo in IB3 rispetto alle C38, e silenziamenti genici di PIASy (specifica E3 ligasi di SUMO) o di SUMO hanno dimostrato uno sbilanciamento a favore della Ubiquitinazione. Tali informazioni hanno suggerito l’ idea che la persistenza di TG fosse Sumo-dipendente. L’inibizione della produzione di ROS in cellule IB3 ha evidenziato che i livelli di SUMO e PIASy erano notevolmente abbassati e con essi la SUMOylazione di TG2. I dati sottolineano che l ambiente pro ossidativo della CF giuda la persistenza di TG2 via SUMOylazione PIASy-dipendente. C) Gli effetti della persistenza di elevati livelli di TG2 mediati da CFTR via ROS hanno dimostrato avere importanti effetti pro-infiammatori che portano ad elevati livelli di IL8. Per individuare le strade più idonee per antagonizzare questa pathway che dal difetto di CFTR conduce all'infiammazione, abbiamo studiato gli effetti della persistenza della TG2 negli epiteli FC. La TG2 induce modifiche posttraslazionali di proteine substrato ed abbiamo identificato un link diretto tra TG2 e il modulatore anti–infiammatorio PPAR͎ (Daynes, R. A., and D. C. Jones. 2002. Nat. Rev. Immunol.). Infatti I livelli di PPAR͎ sono stati ritrovati bassi in IB3 rispetto alle C38. Questa relazione fra TG2 e PPAR͎ è stata dimostrata dipendere dalla attività cross linkante operata da TG su PPAR͎ con formazione di legami isopeptidici. Studi con agonisti o inibitori del PPAR͎ hanno dimostrato che il crosslinking di PPAR͎ e' uno dei meccanismi attraverso cui TG2 induce infiammazione in epiteli FC. D) Studi di microscopia confocale hanno poi evidenziato che il PPAR͎ e' sequestrato in struttre note come aggresomi (Kopito, R. R. 2000. Trends Cell Biol.), che rappresentano stazioni di accumulo di proteine misfolded o modificate post- traslazionalmente e che sono eliminate attravesro un meccanismo alternativo di degradazione, noto come autofagia. Nelle cellule epiteliali FC osserviamo un accumulo di aggresomi, caratterizzati dalla presenza del marcatore HDAC6 (Kawaguchi, Y.,et al. 2003.Cell) e di ubiquitina, con sequestro di PPARg oltre ad aggresomi contenenti la stessa CFTR misfolded. L'accumulo di aggresomi in cellule FC suggerisce una alterazione complessa del degradasoma e suggerisce che le cellule con difettiva CFTR potrebbero presentare un difetto del meccanismo di autofagia.L'autofagia e' un meccanismo che le cellule adottano in condizioni di stress o di carenza nutrizionale come meccanismo di sopravvivenza (Mizushima, N., et al. 2008 Nature; Moreau, K., Luo, S. & Rubinsztein, D. C. 2010 Curr. Opin. Cell Biol.). Con tecniche di trasfezione, rel time PCR, western blot, immunoprecipitazione, microscopia confocale e microscopia elettronica, Noi abbiamo dimostrato che le cellule epiteliali FC presentano un difetto di autofagia con riduzione di autofagosomi, aumento del pool di proteine ubiquitinate e aumento del marcatore p62 che aumenta in maniera significativa in condizioni di difettiva autofagia. Abbiamo poi dimostrato che questo difetto di autofagia e' dovuto al crosslinking di beclin 1 da parte di TG2. Beclin 1 e' una proteina cruciale per la formazione di membrane di autofagosomi (Axe, E. L. et al. 2008, J. Cell Biol.) e la sua associazione nel reticolo endoplasmico (ER) con altre proteine del complesso III del phosphatidylinositol-3-kinase complex (PI3K), e' cruciale per iniziare il processo autofagico (Matsunaga, K. et al.2009,Nature Cell Biol.; Zhong, Y. et al. 2009,Nature Cell Biol.). Noi abbiamo dimostrato che le modifiche posttraslazionali (crosslinking) di beclin 1 operato dalla TG2 induce uno spiazzamento del complesso III PI3K dall'ER e suo sequestro in aggresomi con inibizione dell'autofagia. La overespressione di Beclin 1 cosi' come l'inibizione della TG2 ripristina la normale funzionalità del

CHAPTER I processo autofagico con eliminazione degli aggresomi e riduzione deli marcatori di infiammazione fra cui IL8. E) Pertanto il difetto di CFTR induce una attivaione dll'asse ROS-TG2 che attraverso una inibizione dell'autofagia conduce ad infimmazione. Per dimostrare la valenza di questo meccanismo nella patogenesi dell'infiammazione in FC, abbiamo utilizzato altri 2 modelli biologici. Il primo e' rappresentato dalla coltura ex vivo di mucose di polipo nasale di pazienti con FC, e il secondo è il modello murino di malattia. (Mall, M.et al. 2004, Nature Med.).I topi sono stati trattati con cystamine o NAC. Queso trattamento ha indotto riduzione dell'infiammazione polmonare, riduzione di livelli di MIP-2, l'analogo murino della IL8 e riduzione dei neutrofili intratissutali.Nel loro complesso questi dati dimostrano che inibire la TG2 o ridurre i livelli di ROS e' un approccio estremamente utile in vitro e in vivo per controllare l'infiammazione e i livelli di IL-8. Per ipotizzare un trasferimento dei nostri risultati alla biotecnologia industriale si e' pensato di ipotizzare un device che consentisse lo studio della dinamica del processo di migrazione dei neutrofili in risposta al fattore chemoattrattante IL8 che abbiamo visto giocare un ruolo patogenetico nella infiammazione respiratoria in FC. E’ stata realizzata una matrice tridimensionale di gel di collagene (principale componente delle matrice extracellulare) per mimare le reali condizioni fisiologiche. Sono stati eseguiti esperimenti di motilità su neutrofili in presenza o in assenza di IL8 in una matrice 3D di gel, utilizzando un’appropriata cella di flusso. La matrice di gel è costituita da una soluzione liquida di collagene al quale è aggiunto e una sospensione cellulare contenente 2·105 cell\ml e fatta gelificare per 30minuti in un incubatore per cellule. Per lo studio del fenomeno è stata progettata e realizzata una cella di chemiotassi, costituita da un blocco di alluminio, adeso ad un vetrino portaoggetti, che alloggia un pozzetto al cui centro presenta un blocchetto (inserito tra due guide) sul quale viene fatta aderire una membrana semipermeabile per la realizzazione di un flusso diffusivo. La cella di chemiotassi contenente le cellule in gel di collagene, in assenza o presenza di chemoattraente, viene alloggiata sul tavolo portaoggetto di un sistema time-lapse, il quale consta di un microscopio rovesciato equipaggiato di motorizzazione per il tavolino portaoggetti e di un micro incubatore per cellule. In maniera completamente automatizzata le immagini vengono acquisite a diverse profondità, effettuando delle scansioni 3D del campione in esame ad intervalli regolari nel tempo.Le immagini sono poi analizzate tramite tecniche di image analysis al fine di: determinare la posizione 3D per ogni istante di tempo, ricostruire la traiettoria seguita, e definire la direzionalità perseguita da ogni singola cellula. Per validare il sistema è caratterizzare il flusso diffusivo utilizzando destrano fluorescente (FITC-destrano), avente peso molecolare paragonabile a quello del chemoattraente. In seguito, i primi esperimenti di controllo sono stati condotti su neutrofili (da donatori sani) in presenza e assenza di uno stimolo chemiotattico. I risultati ottenuti indicano, come atteso, un movimento cellulare del tutto casuale in assenza di IL8 e fortemente direzionato in condizioni di presenza dello stimolo. Questo strumento di analisi direzionale e quantitativo potrà consentire innanzitutto di utilizzare questo modello per screening di sostanze con potenziale effetto chemoattrattante per leucociti in diverse patologie umane. La definizione dei meccanismi che regolano la produzione di IL8, come dimostrato nel percorso di dottorato, potrà inoltre aprire la strada ad un nuovo approccio per testare la produzione di chemochine da parte di cellule epiteliali di CF o di altre patologie

CHAPTER I infiammatorie croniche poste nel device insieme ai leucociti per screenare "in situ" l'effetto di inibitori di pathways che regolano la produzione di IL8. I risultati ottenuti nel corso di questo percorso di dottorato di ricerca aprono la strada ad un ulteriore modello di potenziale applicazione alle Biotecnologie Industriali, l'analisi dinamica e direzionale 2D di cellule epiteliali poste in piastre di coltura. Infatti l'attivazione dell'asse ROS-TG2, oltre ad avere un impatto sui meccanismi patogenetici della FC precedentemente descritti, e' anche in grado di influenzare la fisiologia del citoscheletro cellulare, fattore notoriamente cruciale nel mediare l'interaione fra popolazioni cellulari. A tale scopo abbiamo utilizzato un modello di un'altra patologia infiammatoria cronica, la celiachia, in cui precedenti evidenze, fra le quali quelle ottenute dal nostro gruppo di ricerca, hanno dimostrato il ruolo determinante della TG2 nella patogenesi della malattia. La celiachia e' una intolleranza alimentare caratterizzata da una attivazione del sistema immune mucosale a particolari sequenze peptidiche della gliadina, una proteina del frumento. In particolare alcuni peptidi, fra cui quello corrispondente alla sequenza aminoacidica 31-43 della A-gliadina, sono in grado di attivare le cellule epiteliali intestinali predisponendo la mucosa intestinale alla azione di altre frazioni peptidiche, fra cui la 56-68 (a-9) che in un contesto di attivazione della risposta innata sono presentati alle cellule T gliadina-specifiche, evocando in tal modo una attivazione del sistema immune adattativo(Sollid LM.. 2002, Nat Rev Immunol; Sollid LM. 2000 Annu Rev Immunol; Shan L, et al. 2002 Science; Maiuri L, et al. Lancet, 2003).Nel corso del dottorato abbiamo dimostrato come il peptide 31-43 sia internalizzato nelle cellule epiteliali e accumuli nei lisosomi con un ritardo della degradazione che induce un aumento di produzione di ROS e conseguente upregolazione della TG2. Cio' induce un profondo riarrangiamento del citoscheletro con riarrangimento dei filamenti di actina. Per le ragioni precedentemente dette, abbiamo voluto verificare se la tossicità del peptide 31-43 potesse impattare sul movimento e interazione delle cellule epiteliali intestinali poste in piastra di coltura, in considerazione dei suoi effetti sul citoscheletro. A tale scopo sono stati condotti esperimenti in 2D in piastre da coltura su linee cellulari epiteliali di T84 in presenza o meno di stimoli infiammatori, quali il p31-43, seguiti da una regolazione della TG2 mediante inibizione con un anticorpo monoclonale (TG 4G3) o mediante inibizione dei ROS utilizzando l’ EUK134 (mimetico chimico della catalasi- Superossido Dismutasi). I movimenti delle cellule sono stati seguiti mediante la tecnologia Time-Lapse; è stato eseguito un esperimento in 2D in cui le immagini dei campi di vista sono acquisite ogni 15 min, per un totale di 48h. La caratteristica morfologia cellulare delle T84 si esplicita nella formazione di strutture isolari che si accrescono durate le fasi si duplicazione e mostrano movimenti di associazioni di più isole (simile ad un processo di inglobazione). Le immagini ottenute mostrano una chiara alterazione della motilità delle isole in presenza del solo p31-43 rispetto al controllo non trattato (medium alone). L’ aggiunta dei modulatori di TG2 e dei ROS riporta il sistema ad una condizione paragonabile al campione non trattato, mostrando una riduzione di motilità rispetto al p31-43. Le cellule sono state in seguito fissate e utilizzate per l’ osservazione in fluorescenza di un marcatore tipico del citoscheletro, la falloidina; le immagini acquisite tramite microscopia confocale mostrano un chiaro riarrangiamento citoscheletrico nel

CHAPTER I campione trattato con p31-43 rispetto al non trattato; analogamente al movimento, in seguito alle inibizioni della TG2, il citoscheletro riprende una struttura paragonabile al non trattato. I risultati dimostrano quindi che l’ alterato profilo infiammatorio impatta sul movimento cellulare e consente di identificare sostanze capaci di modulare le cause che determinano alterazioni della motilità. Questo modello offre importanti applicazioni potenziali in diverse patologie umane in quanto l'alterazione del citoscheletro e la TG2 sono entrambi coinvolti in processi di migrazione e interazioni cellulari alla base della invasività di cellule neoplastiche.

Conclusioni e prospettive future

I risultati della ricerca svolta nei tre anni di Dottorato offrono innanzitutto un importante contributo alla definizione dei meccanismi patogenetici dell'infiammazione polmonare in FC e identificano nuove opzioni terapeutiche per i pazienti. In particolare i principali risultati dello studio che hanno riscontro in 4 lavori pubblicati su riviste internazionali definiscono: a) il ruolo della TG2 come fattore patogenetico nuovo nella FC b) il ruolo di modifiche posttraslazionali della TG2 come link fra difetto genetico di CFTR ed infiammazione c) i meccanismi attraverso i quali l'asse ROS-TG2 inibisce l'autofagia nelle cellule epiteliali respiratorie FC. Questo studio definisce per la prima volta la FC come malattia correlata ad autofagia. Questo percorso di studio fornisce poi il razionale per l'implementazione di modelli di motilità/migrazione cellulare come strumenti idonei per testare l'efficacia di molecole a potenziale impatto farmacologico. La possibilità di disporre di un modello di analsi quantitativa e qualitativa potrà consentire di saggiare l'efficacia di diversi farmaci che mirano a controllare l'invasività di cellule neoplastiche e per modulare la farmaco-resistenza che costituisce un problema cruciale di terapia. Gli effetti da noi osservati di modulazione di TG2 e di ROS su questi processi dinamici potrà consentire la messa a punto di device idonei sia per finalità di ricerca che per screening dell'efficacia di farmaci. Infine la difficoltà di utilizzare modelli di malattia e' spesso un limite per effettuare screening di sostanze a potenziale impatto terapeutico. Questo modello dimostra di possedere sia i vantaggi di un "modello in vitro", cioe' la manipolabilità e versatilità, sia le caratteristiche peculiari del modello ex-vivo. E' infatti possibile screenare farmaci sul loro target naturale, cioe' le cellule del paziente poste in coltura con il grande vantaggio di poiter tener conto della variabilità interindividuale di risposta al trattamento che costituisce un problema cruciale nella farmacologia clinica. Queste caratteristiche rendono il sistema per l’analisi quantitativa della motilità e delle interazioni cellulari sviluppato nell’ambito di questo lavoro di tesi adatto per applicazioni biotecnologiche industriali.

CHAPTER I

Dati ottenuti durante il periodo di PhD sono inseriti nelle seguenti pubblicazioni:

Luciani A, Villella VR, Esposito S, Brunetti-Pierri N, Medina D, Settembre C, Gavina M, Pulze L, Giardino I, Pettoello-Mantovani M, D'Apolito M, Guido S, Masliah E, Spencer B, Quaratino S, Raia V, Ballabio A, Maiuri L. Defective CFTR induces aggresome formation and lung inflammation in cystic fibrosis through ROS-mediated autophagy inhibition. Nat Cell Biol. 2010,12(9):863-75.

Luciani A, Villella VR, Vasaturo A, Giardino I, Pettoello-Mantovani M, Guido S, Cexus ON, Peake N, Londei M, Quaratino S, Maiuri L. Lysosomal accumulation of gliadin p31-43 peptide induces oxidative stress and tissue transglutaminase-mediated PPARgamma downregulation in intestinal epithelial cells and coeliac mucosa. Gut. 2010; 59(3):311-9.

Luciani A, Villella VR, Vasaturo A, Giardino I, Raia V, Pettoello-Mantovani M, D'Apolito M, Guido S, Leal T, Quaratino S, Maiuri L. SUMOylation of tissue transglutaminase as link between oxidative stress and inflammation. J Immunol. 2009; 183(4):2775-84.

Maiuri L, Luciani A, Giardino I, Raia V, Villella VR, D'Apolito M, Pettoello-Mantovani M, Guido S, Ciacci C, Cimmino M, Cexus ON, Londei M, Quaratino S. Tissue transglutaminase activation modulates inflammation in cystic fibrosis via PPARgamma down-regulation. J Immunol. 2008,180(11):7697-705.

CHAPTER I

ABSTRACT

Cell migration is a very complex mechanism linked to the inflammatory response as well as to several processes of cell biology and development. The movement of cells in response to a chemokine gradient is the most popular form of interaction between cell environment and surface and specific surface receptor triggers an intracellular signaling pathway. This phenomenon is known as chemotaxis and is characteristic of neutrophil movements in response to the chemokine Interleuchin-8 (IL8) in a 3D space. Another paradigm of dynamic interaction between cells is the mechanism underlying the confluency of cells cultured in a Petri dish. This process is particularly studied in growing epithelial cell lines, and is influenced by still undefined external stimuli as well as by cell proliferation and cytoskeleton reorganization, that allow the contact and interaction between cells in 2D space. Chronic inflammation is an ideal condition for the study of cell migration and its deregulation; understanding the mechanisms involved in cell motility and their putative modulation, is the first step in setting new appropriate models of study. During PhD project, different cellular models have been used: IB3 cell lines belonging to patients with Cystic Fibrosis (caused by mutations of cystic fibrosis transmembrane regulator (CFTR)), and T84 cell lines, a popular model of study of Coeliac Disease, a common intolerance to proteins of wheat. Both disease are characterized by a pro-inflammatory milieu with high level of tissue Transglutaminase (TG2), a multi-functions enzyme with a defined role in several human pathologies. Cystic Fibrosis is the prototype of diseases in which an uncontrolled production of IL8 leads to a dysregulated neutrophil recruitment. The IB3 cell lines have been used to understand a molecular mechanisms of enhanced IL8 production and their modulation. The results evidence i) the role of TG2 as a new pathogenic factor in Cystic Fibrosis; ii) the role of post-translational modifications of TG2 as a link between genetic defect of CFTR ad inflammation iii) the mechanisms of autophagy inhibition via ROS-TG2 axis, in epithelial airway cell line and mice model of CF. The consequence is a fine modulation of IL8 production. The system allows the design of a 3D model to study cell migration under IL8 diffusive flux in qualitative and quantitative ways. Another important pathological system with chronic inflammation is celiac disease and the analysis of interactions of alimentary peptides with the frontline gut epithelium is a useful model of study. The epithelial cells are pivot in the innate immune activation in CD and cytoskeleton rearrangement is the earliest event in such a response to gliadin peptides. The data in T84 cell model show a gliadin peptide-driven pro-inflammatory environment as a consequence of its impaired lysosomal degradation. Moreover, alterations of motility and cytoskeletal reorganization emphasize the “toxic” effect of peptide. Therefore, CD offers an ideal opportunity to set up a 2D model of study of cell motility. Understanding the mechanisms of migration and the identification of appropriate target of modulation of cell recruitment could allow wide industrial applications in biotechnology and will be successful to provide a useful tool to test potential therapeuthic molecules.

CHAPTER I

INTRODUCTION

Motility and Inflammation

Cellular interaction and migration are common physiological events triggered and finely tuned by different mechanisms either intrinsic to the cell or present in cellular environment. Multiple interactions between environment and cell surface induce dynamic modifications of cellular phenotype (1). Cell migration, both single and collective, is a highly integrated multistep process that is essential in embryonic morphogenesis, tissue homeostasis and immune surveillance to inflammation. In collective migration a movement of cohesive groups of cells is required (2) and the single migrating cell shows a highly polarization with complex regulatory pathways that are spatiotemporally controlled (1). Migration contributes to several important pathological processes, including cancer progression and metastasis formation. Metastasis, dissemination of malignant tumors to a distant organ, is the major cause of cancer mortality. Tumor cell motility is the hallmark of invasion and is an essential step in metastasis(3). Cell migration is a very complex phenomenon also linked to the inflammatory response as well as to several processes of cell biology and development. Cell movements into solid surfaces require a controlled sequence of cell protrusions and retractions that mainly depends on sophisticated regulation of the actin cytoskeleton and the contribution of microtubules (4,5). The movement of cells in response to a chemokine gradient is the most popular form of interaction between cell environment and surface in which specific surface receptor triggers an intracellular signaling pathway. The final event is the movement of cells in direction of chemoattractant (6). This process is characteristic of neutrophils movements in response to the chemokine Interleuchin-8 (IL8) in a 3D space. Another paradigm of dynamic interaction between cells is the mechanism underlying the confluency of cells cultured in a Petri dish. This may be considered the prototype of movements in 2D space. This process is particularly studied in growing epithelial cell line, and is influenced by still undefined external stimuli as well as by cell proliferation and cytoskeleton reorganization, that allow the contact and interaction between cells. Cell contact and interaction may in turn impact on cell differentiation, as in confluent intestinal epithelial Caco-2 cell lines, which undergo differentiation with surface expression of brush border enzymes, markers of terminal differentiation. Both attraction forces that compromise the adhesion of cell to fiber of matrix and cellular adaptative mechanisms are required to respond to matrix resistance. In 3-D ECM, in contrast to 2-D substrate, the cell shape is mostly bipolar and the cytoskeletal organization is less stringent, frequently lacking discrete focal contacts and stress fibers.(7) Cell migration can be seen as a cyclic process usually driven by several factors, among the others actin polymerisation, stabilized upon adhesion to the extracellular matrix (ECM). Although matrix adhesion dynamic is very important for understanding cell migration behavior, the molecular mechanisms that regulate adhesion dynamics and signalling in disease conditions are still unclear (8,9). All together these findings indicate that cell motility and migration occur as a consequence of a finely tuned interaction between intracellular and cell surface signaling pathways which can be influenced by a wide range of environmental 10

CHAPTER I factors. Moreover such a signaling cascade may be disrupted in several disease states and migration itself can be considered a key step in disease evolution (10,11). In the complex overview of motility and interaction between the cells within a specific environment, the research of factors influencing the interaction in human pathology is mandatory. As a matter of fact, the progress of many chronic disease is influenced by alterations of motility or interaction. An important instance is represented by neo- plastic movements of cancer cell and its deregulation during disease progress; understanding the mechanism of modulation of factor involved in this deregulation will be relevant for therapeutic applications. The study of “moviments” and their directionality in human pathology is not only relevant for cell-to-cell interactions within a specific environment. Such a process is also a feature of intra-cellular processes where cytoplasmatic organelles, proteins aggregates and proteins (nuclear translocation) move within the cell with defined rules. A typical example is the trafficking and reciclying of membrane proteins within endosome-structure or degradation route (proteasome). These movements are the basis of several processes, among the others autophagy and cytoskeleton rearrangement that in turn can influence the “external” motility. Cell migration may also take place within a proper tissue environment. This is the case of mucosal tissues approaching a luminal cavity, as airways or gut mucosa. In this case the spatial organization of tissues allow a compartmentalization of different cell types whose location reflects a specific function. Epithelial cells are the front line of airways and intestine and represent a set of specialized cells. These cells are also able to orchestrate cell migration within the tissues to maintain a proper tissue homeostasis in response to environmental triggers. In this context they are able to fight "toxic" environmental challenges, either bacterial or alimentary triggers, by secreting specific chemoattractant factors that recruit immunocompetent cells to the "damaged" luminal surface. This process is a well orchestrating event that involves different types of cells moving within the tissue and interacting each other in response to a given triggering factor. The release of chemoattractants is the results of the activation of complex intracellular signaling cascades also involving a cytoskeleton rearrangement and generating a "stressed" environment (5). The physiological mechanisms underlying such a complex production of chemoattractant factors is often impaired in disease conditions in which an improper activation of specific pathways or the lack of physiological mechanisms of control, lead to impaired production of chemoattractants with derangement of cell motylity and physiology. The most popular examples of intra-tissue migration of cells are chronic inflammatory diseases occurring at the mucosal surface. Airway and gut chronic inflammatory diseases are the ideal condition for the study of cell migration and its de-regulation within a controlled environment. In these conditions several cell populations are recruited in response to the enhanced local release of chemo-attractant molecules (chemokines) that favor the infiltration of circulating leukocytes within a disease- target tissue. This phenomena is noted as chemotaxis and the chemical mediators (chemokine, citokine) recall the white cells from blood; the cells must go through different layers(blood vessel, extracellular matrix, epithelia) (12). The formation of chemical gradient (cytokine-chemokine) become an important motor for cell migration. One of main factors regulating the chemokine-chemical gradient is the presence of an inflammatory status. Inflammation is a physiological process of response to tissue

CHAPTER I injury. However, the dysregulation of the proper control of inflammation, induces a massive infiltration of leukocytes (neutrophils, monocytes, lymphocytes) within the target tissue as it occurs in gut, airways, skin, joints leading to chronic inflammatory response and tissue damage. Inflammation is also strictly linked to many events that favor cell motility, such as cytoskeleton rearrangement In chronic inflammatory diseases at the mucosal sites the rules of cell migration are strictly dependent on disease pathogenesis as well as on the nature of the unwanted triggering event dictating the time-course and the type of the release of chemoattractants by the injured epithelia (13). In arways, bacterial challenge is a common "toxic" event which requires neutrophil recruitment. As a consequence, the release of IL8, a potent neutrophil chemoattractant, is required. The prototype of diseases in which an uncontrolled production of IL8 leads to a dysregulated neutrophil recruitment, is Cystic Fibrosis, a most common lethal genetic disorder in caucasians characterized by chronic inflammation and recurrent pulmonary bacterial infections. These evidences highlight that the identification of factors that dysregulate the complex mechanism of motility is the key model to develop molecules able to regulate this factor in order to find putative biotechnological applications. The choice of a suitable disease model for the study of potential factors involved in motility\interaction processes, must take in account that these mechanisms are a pathogenic element of disease and not only a bystander event. Understanding the specific pathways involved may lead to the identification of molecules able to modulate this biological process in human pathology. During PhD project, different cellular models of these two main diseases have been used as model of study to understand the molecular mechanisms that influence inflammation-triggered cell movements : IB3 and C38 cell lines belonging to patients with Cystic Fibrosis and T84 cell line, a popular model of study of Coeliac Disease, a common intolerance to proteins of wheat. Both disease are characterized by a pro- inflammatory milieu with high level of tissue Transglutaminase (TG2), a multi- functions enzyme with multiple functions, as cross-linking of substrate protein, isomerase or G protein activity, with a defined role in several human pathologies (33). Cystic fibrosis is caused by mutations of cystic fibrosis transmembrane regulator (CFTR), which functions as a chloride channel in epithelial membranes. More than 1000 mutations in CFTR gene have been identified and they can be grouped into six classes: (1) CFTR is not synthesized; (2) defective processing; (3) regulation; (4) conductance; (5) partial defective production or processing; or (6) defective regulation. The most common CFTR mutation belongs to class 2, and is caused by the deletion of phenylalanine at position 508 (F508del) of CFTR. The impaired chloride transport leads to water/volume depleted periciliary liquid with a increase of mucus viscosity . Bacteria invading the cystic fibrosis lung are trapped within this viscous mucus layer at the surface epithelial layer. As a consequence, recurrent pulmonary infections, principally by Pseudomonas Aeruginosa, colonization together with chronic inflammation are main features of CF lung pathology(14). CF chronic inflammation is neutrophil-mediated, as circulating neutrophils are recruited within the airways and contribute to the extensive lung damage and to the development of impaired lung function. Accordingly, CF lung pathology is associated with a marked increase of proinflammatory cytokines, such as TNF--6, IL-!- 17 (15, 16), and the potent neutrophil chemoattractant and activator IL-8, which recruits large numbers of neutrophils into the airways.

CHAPTER I

However, conflicting results have been reported on whether CF airways undergo proinflammatory response in the absence of bacterial infections (17-19). Several studies have shown that both cytokine profile secretion and NF-'B activation are similar in CF and normal cells (20, 21). However, mounting evidences suggest that inflammation may occur before infection, and CFTR-defective cells have an intrinsically proinflammatory phenotype (22-24). Recently, it has been shown that CFTR is a negative regulator of NF-'B-mediated innate immune response, and its localization to lipid rafts is involved in control of inflammation (25). The neutrophil involvement in CF chronic airway inflammation could suggest a development of a 3D motility model to study neutrophils migration under diffusive flux of IL8 and their containment upon identification of the main rules governing IL-8 production. As chronic inflammation may represent a driving force of cell motility/migration, we could envisage several disease-related models of study, either 3D or 2D models. The study of cell migration in a given disease-related model could help in defining novel devices useful in validating several attracting mechanisms and related molecules with potential implicatrions in human therapy. Among the others, another important pathological system with chronic inflammation is celiac disease and the analysis of interactions of alimentary peptides with the frontline gut epithelium is a useful model of study (a popular model of study is represented by the T84 cell lines). Coeliac disease (CD) is a very common pathology that affects 1% of the general population, caused by the ingestion of gluten proteins,in susceptible individuals (26- 28). Its strong associations with HLA class II (HLA DQ2 or DQ8) and a well-defined triggering antigen suggest that coeliac disease is a prototype of T-cell-mediated disease.(26). Evidence suggests that CD4(+) T cells are central in controlling the immune response to gluten that causes the immunopathology (29-31). The primary HLA association in the majority of CD patients is with DQ2 and in the minority of patients with DQ8. DQ2 or DQ8 are the predominant restriction elements for gluten- specific T cells. Lesion-derived T cells predominantly recognize deamidated gluten peptides. A number of distinct T cell epitopes within gluten exist. DQ2 and DQ8 bind the epitopes so that the glutamic acid residues created by deamidation are accommodated in pockets that have a preference for negatively charged side chains. Evidence indicates that deamidation in vivo is mediated by the enzyme tissue transglutaminase (TG2). This may result in the formation of complexes of gluten- TG2. These complexes may permit gluten-reactive T cells to provide help to TG2- specific B cells by a mechanism of intramolecular help, thereby explaining the occurrence of gluten-dependent TG2 autoantibodies that is a characteristic feature of active CD(27). Moreover it is well known that a co-operation between the innate and adaptive branches of the immune response are crucial in CD pathogenesis. Some gliadin peptides, which are not recognized by gliadin-specific T cells, are effective in activating the innate response either at epithelial or dendritic cell level. This innate activation is essential in setting the tone of the adaptive response, as the immunodominant peptides turn on toxic only if the gut mucosa has been primed by the so-called "innate" gliadin fractions (77). These peptides are able to induce profound epithelial changes in celiac mucosa as well as in models of CD, as T84 or Caco2 cell lines. Innate epithelial activation include phosphirylation of proteins, cytoskeleton riorganization and even IL-15-mediated apoptosis (77). The peptide p31-43 is the prototype of these peptides and it has been reported as also effective in inducing early TG2 activation.

CHAPTER I

A link between these two models of human diseases, CF and CD, the first monogenic, the second one multifactorial with a strong involvement of the immune response, is the presence of high tissue levels of the enzyme Tissue Transglutaminase (TG2) a protein with pleiotropic functions, as a consequence of the genetic defect in CF, or triggered by gliadin peptides in CD. In addition, both diseases show a chronic inflammatory phenotype, although neutrophil-mediated in CF and devoid of neutrophil component in CD. Finally in these two pathological conditions surface epithelial cells are key players in disease pathogenesis. The epithelial cells are pivot in the innate immune activation in CD and cytoskeleton rearrangement is the earliest event in such a response to gliadin peptides; moreover cytoskeletal reorganization is a known factor of cell differentiation (32) . Therefore, CD offers an ideal opportunity to set up a 2D model of study of cell motility. In this study we have also developed a 2D motility model to test i) the effects of gliadin peptides in deranging cell motility and interaction in a 2D monolayer epithelial culture model and ii) to test the role of several modulatory molecules with potential therapeutic application.

CHAPTER I

AIMS

Cell migration is a very complex phenomenon that regulates the inflammatory response as well as several processes of cell biology and development. The identification of a specific regulatory pathway, offers a potential strategy to modulate many human pathology. Understanding the mechanisms involved in cell motility and their putative modulation, is the first step in setting the new appropriate model of study. Therefore, the PhD activity consisted of two main activities: A) the study of the molecular mechanisms leading i) to release and modulation of major chemoattractant factors’ (such as IL8) and their containment; ii) the effects of chronic inflammation in deranging cell-to-cell interaction and the putative mechanisms of control. To this aim two human disease models of chronic inflammation, Cystis Fibrosis with IL8-driven neutrophil-dominated chronic airway inflammation and iii) Celiac disease, the prototype of food induced activation of the mucosal immune response have been used. B) the development of motility models: i) a 3D model of neutrophil migration in an IL- 8-rich environment; ii) a 2D model of interaction of epithelial cells upon challenge with celiac disease-triggering peptide in presence or absence of putative modulator factors. Both disease models are characterized by increased tissue levels of TG2 and chronic inflammation, as constitutive chronic airway inflammation characterizes Cystic Fibrosis, and gluten-triggered gut inflammation is a feature of Celiac Disease. The first condition is characterized by a defective CFTR-driven pro- inflammatory environment mediated by increased levels of Reactive Oxygen Species (ROS) and high level of TG2, a pleiotropic enzyme involved in many pro- inflammatory conditions. In such a disease model the modulation of ROS-TG2 axis leads to restoration of cellular homeostasis with inhibition of IL-8 production. To understand the mechanisms of TG2 regulation, we use (i) classical and innovative technique of cellular and molecular biology in CF and control epithelial cell lines; ii) an “ex vivo” model of human nasal biopsies; iii) “in vivo” mice models of disease. These in vitro and in vivo approaches allow the identification of the link between defective CFTR and inflammation through the disregulation of TG2-ROS axis. The second condition, Celiac Disease, allow the identification of the rules of TG2 upregulation upon challenge with specific gliadin-derived triggers. In such a condition the study revealed how the modulation of ROS-TG2 controls gliadin- induced mucosal inflammation. To translate these results into an appropriate study model of motility for putative biotechnological applications, a 3D model for neutrophils migration assay has been set up. The 3D model is an ideal model to study the movements of cells in a 3D matrix of gel, where a diffusive flux of molecule induces migration. Moreover a 2D model has been used to test whether controlling ROS-TG2 pathway might restore the derangement of cell-to-cell interaction and cytoskeleton reorganization taking place upon challenge of T84 intestinal epithelial cells with gliadin-derived peptides. The results highlight the feasibility of this model as a prototype to screen, in a very fast way, different molecules with potential therapeutic application.

CHAPTER II

RESULTS

The first part of PhD research program focused on the analysis of the molecular mechanisms responsible for the induction of the pro-inflammatory phenotype that characterizes Cystic Fibrosis airways, and the processes that regulate high TG2 levels. Transglutaminase 2 (TG2) is an inducible transamidating acyltransferase that catalyzes Ca(2+)-dependent protein modifications. It acts as a G protein in transmembrane signalling and as a cell surface adhesion mediator, this distinguishes it from other members of the transglutaminase family (30 ). In mammals, nine distinct TGase isoenzymes have been identified at the genomic level; however, only six have so far been isolated and characterized at the protein level. All members of this superfamily possess a catalytic triad of Cys-His-Asp or Cys-His-Asn. Recent structural studies have described the regulatory mechanism of the transamidating activity for the TG2, but can be assumed to act in all the family members. The arrangement of the amino acids of the catalytic centre (Cys277, His 335 and Asp358) in a charge-relay catalytic triad, analogous to that of thiol proteinases such as papain, confers high reactivity on Cys277 to form thioesters with peptidylglutamine moieties in the protein substrate. This high reactivity, in the presence of Ca2þ, of Cys277 has been employed to develop a wide range of active-sitedirected irreversible inhibitors of the enzyme(34) The sequence motifs and domains revealed in the recent TG2 structure, can each be assigned distinct cellular functions, including the regulation of cytoskeleton, cell adhesion and cell death. Ablation of TG2 in mice results in impaired wound healing, autoimmunity and diabetes, reflecting the number and variety of TG2 functions. An important role for the enzyme in the pathogenesis of coeliac disease, fibrosis and neurodegenerative disorders has also been demonstrated, making TG2 an important therapeutic target. Understanding the mechanisms of migration and the identification of appropriate target of modulation of cell recruitment could allow wide industrial application in biotechnology.

1.The Inflammatory environment of Cystic Fibrosis

1.1.The up-regulation of TG2 in human CFTR-defective cells

To underpin the molecular link between a defected CFTR and the excessive inflammatory responses typical of CF airways, here we have studied CFTR-defective bronchial epithelial cell lines, IB3-1 human CF bronchial epithelial, (carrying F508del/W1282X CFTR mutation) and isogenic stably rescued C38 . We demonstrated in CF airway epithelial cell lines and in human CF nasal biopsies increased levels of TG2. We observed levels of TG2 protein in IB3 and C38 cells, that were analysed by Western Blot. The figure 1A shows high levels of TG2 (72 kDa) in IB3 compared to

CHAPTER II

FIGURE 1. TG2 in CF cell line

CHAPTER II

C38. Therefore, we explored whether ,besides TG2 high levels, also its enzymatic activity were up-regulated in CFTR-defective cell lines. The confocal microscopy images reveled a signi[ C38, upon incubation with biotinylated monodansylcadaverine (bio-MDC; Molecular Probes), a known molecule substrate of TG2 (Fig. 1A). To assess that the pathway described to date was a direct consequence of CFTR dysfunction, we blocked the functional CFTR in the normal bronchial 16HBE cell line with the selective inhibitor CFTRinh-172 (22). The enzymatic activity of TG2 and protein were high increased in presence of CFTRinh-172 compared to control without inhibitor, as shown in confocal microscopy images in Fig. 1B. These results clearly indicate that CFTR-defective epithelial cells are characterized by an intrinsic TG2 functional increase. Exposure of the IB3-1 cells to the irreversible TG2-speci[ !"#$ % M. Grif[ & ' % '()(*$+-37) as well as reversible inhibitor cystamine or gene silencing by TG specific siRNA, reveled that TG2 activity also in\ - \ ( !tion of TG2 induced a signi[ reduction of the phosphorylated p42/p44 MAPKs, of the nuclear translocation of NF- ' /01- -1 cells(Fig.1C).

TG2 is therefore a master regulator of inflammation in CF airways and mediates the production of inflammatory cytokines and chemokines via targeting key regulatory pathways of inflammation.

1.2. SUMOylation of Tissue Transglutaminase

Because TG2 activity is also regulated by ROS (38) we tested whether ROS in\ / -induced in\ n CF. In agreement with previous reports, high levels of ROS were only detected in CF tissues and CFTR-defective cells (data not shown) (39). The IB3 cells were treated with ROS scavenger NAC (N-Acetil- Cysteine) or EUK 134 and analysis by confocal microscopy and ELISA assay shown a signi[ / 2 / 23433 2 2 $/ not shown), and TNF-a release (Fig. 2A) in both CF tissues and IB3-1 cells. The inhibitor CFTRinh-172 in 16_HBE normal bronchial cell lines reveled an increase of intracellular ROS (Fig. 2B) and increase of p42/44 phosphorylation. Altogether, these results clearly demonstrated how pro-oxidative environment influences TG persistence and CFTR disruption directly leads to the development of airway in\ 51( TG2 is regulated by retinoids, steroid hormones, peptide growth factors and cytokines that also lead to a time-dependent decrease in TG2 ubiquitination (40). This indicates that posttranslational control mechanisms may regulate TG2 tissue levels in CF airways. We focused on small ubiquitin like-modifier (SUMO) posttranslational modification since this has been defined as a central way of regulating key cellular functions and stability of proteins (41,42). SUMOylation has been defined as a key player of the post-translational network to regulate key cellular functions including transcription, nuclear translocation, stress response and chromatin structure as well as of diversifying localization and even stability of the modified proteins (41,42). Sumoylation is accomplished via an enzymatic cascade involving, among the others, E3 ligases, that catalyze the transfer of SUMO from the

CHAPTER II

FIGURE 2. CFTR inhibition induces airways inflammation

conjugating enzyme UBC9 to a substrate (41). E3 ligases have gained a central role in the SUMO machinery, since they regulate sumoylation in response to different stresses (41). To investigate whether posttranslational modifications could result in the persistence of high levels of TG2 protein, the protein extract of IB3-1 and C38 cells were analyzed by Western blots and revealed that SUMO-1 protein level was increased in IB3-1 cells as compared with C38 cell lines (Fig. 3A). Sequence analysis and SUMO motif screening revealed three SUMO accepto$17-89*$3* : ( The experiments of figure 3A shows, also, that when TG2 immunoprecipitates from IB3-1 cells were probed with the anti-TG2 Antibody, two TG2 bands were detected, with the upper band corresponding to the SUMOylated TG2. Since Protein Inhibitor of Activated STAT (PIAS)y, a member of the PIAS family (43), has recently been defined as the first SUMO ligase for NF-kB essential modulator (NEMO) (43) and PIASy-NEMO interaction is mediated by Reactive Oxygen Species (ROS) (43), we also investigated whether PIASy-TG2 interaction could mediate ROS-driven post-translational modifications of TG2. To investigate whether PIASy could mediate TG2 SUMOylation, we performed experiment of Immunoprecipitation of PIASy and TG2 in IB3-1 cells . IB3 cells were treated or not with virus infection of human manganese superoxide dismutase (MnSOD) to regulate the pro-oxidative environment of cell. The IP analysis underline a direct interaction between PIASy and TG2 shown with band corresponding to reciprocal antibodies of IP, with a strong decrease of interaction in presence of MnSOD; as well as PIASy-TG2 interaction also TG2 Sumoylation is decreased upon MnSOD infection (Fig 3B). The western blot analysis of SUMO, TG2, PIASy evidenced the reduction of level with MnSOD(Fig.3C). The data of experiments indicate that in CF airway epithelia the pro-oxidative

CHAPTER II

FIGURE 3. PIASy mediates TG2 SUMOylation in CF airway epithelial cells intracellular milieu leads to PIASy-mediated TG2 sumoylation. SUMOylation influence ubiquitination since these two main posttranslational changes may compete each other for the same lysine residues on the aminoacid sequence. SUMOylation may induce protein stabilization by blocking ubiquitination of the same lysine residues (44). We performed experiments of SUMO or PIASy gene silencing. IP studies revealed an increase of TG2 ubiquitination upon proteasome inhibition by MG132 (Fig.4A,B), thus allowing TG2 to be targeted to proteasome for degradation. This induced decreases of TG2 protein (Fig. 4C) and activity. Therefore, oxidative stress increases PIASy protein levels and favors TG2 SUMOylation that leads to the persistence of high TG2 tissue levels by downregulating TG2 ubiquitination and proteasome degradation. 20

CHAPTER II

FIGURE 4. Decrease of TG2 protein levels and TG2 activity by either SUMO or PIASy gene silencing

CHAPTER II

TG2 SUMOylation might provide the missing link between cellular stress and inflammation. We investigated whether the control of TG2 SUMOylation might modulate TG2-driven inflammation (24). The figure 5 demonstrated that gene silencing of either PIASy or SUMO by specific siRNAs induced a significant decrease of p42\44 phosphorylation(Fig.5A) and TNF- $1%(+B) in IB3-1 cells.

FIGURE 5. SUMO-1 or PIASy gene silencing control inflammation in CF airway epithelial cells

CHAPTER II

1.3.Imapct of TG2 persistence in CF epithelia

The data underline that CF environment ROS-mediated leads to a persistence of TG2 with strong pro-inflammatory effects and increase of citochine release sucs as 01 /#( To identify a more effective way to antagonize this pro-inflammatory pathway CFTR-driven, we studied the effect of TG2 persistence in CF epithelia. TG2 is an enzyme with a vast array of biological functions that mainly catalyzes cross-links or deamidation of target proteins in the presence of high Ca2+ levels. Previously, we shown high TG2 protein level in IB3 cell lines compared to low livel in C38. In these data we demonstrated low levels of 55 kDa Peroxisome Proliferator- Activated Receptor gamma (PPAR͎), (Fig 6A) a nuclear hormone receptor expressed in monocytes, macrophages and epithelial cells that negatively regulates inflammatory gene expression by “transrepressing” inflammatory responses (45.). To unravel the putative link between TG2 and PPAR͎, we then performed experiment of Immunoprecipitation (IP) of protein lysates of IB3 vs C38 cell line The IP analysis underline a direct interaction between PPAR͎ and TG2 shown with band corresponding to reciprocal antibodies of IP, with a strong decrease of interaction in presence of TG2 inhibitor (data not shown). We, then tested the hypothesis that TG2 might induce posttranslational modi[ $ -linking) of PPAR͎ in CFTR-defective cells, because the QP and QXXP motifs in the PPAR͎ sequence could be recognized as speci[ activity (40) We demonstrated that in IB3 cell TG2 catalizes crosslinking and ubiquitination of PPAR͎. After IP of PPAR͎ ! ; !/0$"-L-glutamyl)-L-lysine isopeptide antibody that reveled the specific bound between TG2 and target protein. The figure 6A shows high molecular weight corresponding to modification induced by TG2 (130-250 kDa). Exposure of the IB3-1 cells to the irreversible TG2-speci[ inhibitor R283 or silencing of TG2, induced a signi[ / % mass PPAR͎ ,TG2-induced (Fig. 6B) and a signi[ ++ - kDa PPAR͎ in the IB3-1 cells (Fig. 6B). This indicated that the crosslinked PPARy was due to increased TG2 activity.

CHAPTER II

FIGURE 6. Effects of TG2 on PPAR protein expression and localization

In IB3-1 cells, PPAR͎ nuclear translocation was promoted only by R283, whereas rosiglitazone, a PPAR͎ agonist with potent antiin\ / (46), induced only a marginal nuclear PPAR͎ increase (Fig. 6C). In contrast, rosiglitazone alone was able to induce an impressive PPAR͎ nuclear translocation in C38 cells (Fig. 6C). These results indicate that crosslinking of PPAR͎ by TG2 in CFTR-defective cells flavored a pro-inflammatory environment. 24

CHAPTER II

As previously described, TG2 SUMOylation may have great relevance in driving CF inflammatory phenotype and release of pro-inflammatory cytokines and chemoattractants. PPAR͎ which may be targeted by TG2 to crosslinking, may also be targeted by SUMO-1 and undergo sumoylation in response to a PPAR͎ agonists, such as Rosiglitazone (47). Sumoylated PPAR͎ interacts with the nuclear- receptor co-repressor (N-CoR)-histone deacetylase 3 (HDAC3) complex and thereby blocks its ubiquitination, thus maintaining a repressor condition (47). This study shows that sustained TG2 activation inhibits PPAR͎ SUMOylation and its interaction with the N-CoR (47) thus favoring inflammation. Moreover TG2 induces crosslinking and degradation of Ik- - ; ! $3#* / % regulator of NF-kB activation, inhibits Ik->'?@ / 01-kB activation (Figure 7). Immuprecipitation experiments shown an increased PPAR͎ sumoylation (Fig.7A) , blocking TG2 through specific gene silencing or with specific inhibitor cystamine , in response to rosiglitazone (agonist of PPAR͎) in IB3 cell line, ; confocal microscopy images shown an enhanced N-CoR protein and its nuclear localization (Fig. 7B), and favored N-CoR- PPAR͎ interaction (data not shown). Moreover siRNA of TG2 or cystamine induced increase of Ik- /2 ($%(A5*( TG2 SUMOylation may therefore switch off the intracellular regulatory machinery, preventing PPARy SUMOylation, IkBa stabilization, and leading to an uncontrolled in\ammatory response.

CHAPTER II

FIGURE 7. TG2 inhibition modulates PPARt pathways in CF airway epithelia.

CHAPTER II

1.4. Aggresome: a feature of CF epithelia

To test whether persistence of PPAR͎ due to TG2-crosslinkin influenced its proteasome degradation, we performed immuprecipitation of PPAR͎ upon treatments with MG132 (chemical inhibitor of proteasome), and reveled for Ubiquitin. The figure 8A shows that IB3 cell after treatments with MG132 increased PPAR͎ protein level and its ubiquitination and microscopy images evidenced that it was mainly limited to perinuclear localization in CFTR-defective IB3-1 cell line . This perinuclear localization was suggestive of aggresomes, a speci[c cellular response against accumulation of ubiquitinated misfolded and/or aggregated proteins, which might also result from impaired proteasome function (49). Although aggresomes containing PPAR͎ have not been described before, aggresomes containing ubiquitinated misfolded MF508 CFTR molecules have been previously observed as result of overexpression of the defective CFTR (49). The aggresomes of CFTR also contain HDAC6, a microtubule-associated deacetylase that interacts with ubiquitin and stabilizes polyubiquitin chains (50). We have shown by confocal microscopy images that in the CFTR-defective IB3-1 cell line, the perinuclear aggregates of PPAR͎ also colocalize with both HDAC6 and ubiquitin (Fig. 8B); inhibition of proteasome by MG132 increased this colocalization and the level of HDAC6 and Ubiquitin (Fig.8B). The data thus demonstrated the aggresome nature of these aggregates.

FIGURE 8. Aggresome: a feature of CF epithelia

CHAPTER II

1.5. Aggresome and autophagy

In previous data we described the presence of peroxisome proliferator activated receptor (PPAR)-͎ in aggresomes in CF airways (24), and the mutant CFTRF508del has been described as the prototype of a misfolded aggresome-prone protein (51-53) Sequestration of the anti-inflammatory PPAR-͎ into aggresomes, mediated by the increased TG2 levels, contributes to the lung inflammation in CF (54). This is the result of ROS-mediated TG2 SUMOylation inducing the crosslinking, ubiquitylation and aggresome sequestration of several target proteins (54). Because oxidative stress and accumulation of protein aggregates are features of CF airway epithelia (24,54,55), we speculated that the accumulation of aggresomes in CF is due to defective autophagy. Autophagy is a cytoprotective mechanism for the degradation of misfolded/ polyubiquitylated proteins and damaged organelles through lysosome mediated self- digestion (56,57). Autophagy is important in clearing protein aggregates after overload of polyubiquitylated proteins (58,59). Here we demonstrate how defective CFTR drives autophagy inhibition in mouse and human CF airway epithelia by means of the ROS– TG2 pathway, with sequestration of beclin 1 interactome, a key player of autophagosome formation (60-62), in the aggresomes. Rescuing beclin 1 and autophagy restores CFTR trafficking and damps down inflammation. We cultured human CF airway epithelial IB3-1 or CFBE41o- cells(F508del/F508del), C38, or bronchial epithelial 16HBE14o-cells under nutrient starvation to stimulate autophagy (63,64). In first analysis we observed, through transmission electron microscopy, the presence of autophagosomes containing mitochondria and cellular organelles in C38 but rarely in IB3-1 cells (Fig. 9A). The western blot analysis shown the accumulation of LC3, a marker of autophagosomes (64) in IB3 and C38 cells line. Increase of LC3 II, the lipidated product of LC3 (63,64), was observed in C38 but not in IB3-1 cells under starvation (Fig. 9B). E64d and pepstatin A, inhibitors of the lysosomal proteases (Fig. 9C), as well as chloroquine, a weak base disrupting lysosomal functions failed to increase LC3-II, thus indicating that reduced LC3 II levels were not due to LC3 lysosomal degradation. Defective autophagy in IB3-1 cells was also revealed by the increased pool of polyubiquitinated proteins, and accumulation of p62 (Fig. 9C,D), a LC3- and ubiquitin-binding protein, that accumulates in intracellular aggregates under defective autophagy (63,64). IB3-1 cells were also resistant to rapamycin, a known autophagy inducer (63,64). Real-time PCR analysis of several autophagy-related genes(64,65), ULK1, ULK2, ATG5, ATG6 (beclin 1), ATG7, ATG14, Bcl-2, vacuolar protein sorting (hVps)34, regulatory myrisoylated kinase hVps15 and Ambra1 did not reveal reduced mRNA expression levels in IB3-1 and CFBE41o- as compared to C38 and 16HBE14o- cells(Fig.10A). Growing evidences indicate that beclin 1, a tumor suppressor gene (60,61), interacts with the class III PI3K, hVps34 (65,66) and ER associated class III PI3K activity is crucial in the initiation of autophagosome formation (67,68). Moreover the dissociation of beclin 1 from Bcl-2 promotes autophagy during stress conditions, such as starvation (67-70). PI3K core complex binds either to Atg14L or UV-irradiation-resistance-associated gene (UVRAG) (65,71), involved in distinct functional complexes driving either

CHAPTER II

FIGURE 9. Autophagy is defective in human and mice CF airway epithelia

CHAPTER II autophagy or endosome-to-Golgi retrograde trafficking (71), and Atg14L appears to divert hVps/Class III PI3K into an autophagic role (71) We preformed immunoprecipitation of beclin 1 in IB3-1 and C38 cells with or without starvation and reduced Bcl-2 immunoreactivity was observed in IB3-1 with starvation compared to nutrient condition; hVps34, hVps15, Ambra1 and Atg14L immunoreactivities were detected in beclin 1 immunoprecipitates from C38 but also IB3-1 cells and increased upon starvation, thus suggesting an intracellular environment favourable to autophagy induction (Fig.10B). Reduced protein levels of beclin 1, hVps34, hVps15, Ambra1 and Atg14L (Fig.10B), as well as of UVRAG (data not shown), were observed in IB3-1 cells. Western blots of the insoluble protein fraction revealed increased beclin 1 in IB3-1 as compared to C38 cells, this suggesting retention of beclin 1 as an insoluble protein within aggregates in CF epithelia (Fig.10C). Interestingly, hVps34 (Fig. 10D), beclin 1 and the other beclin 1 interacting proteins (data not shown) colocalized with the ER marker calnexin in C38 but not in IB3-1 cells. In CF cells beclin 1 and beclin 1 interacting proteins, are sequestered in perinuclear aggregates, where they co-localized with the aggresome marker HDAC6 (50) (Fig. 10D). The results indicated an essential role of Beclin 1 in autophagy machinery due to aggresome formation. To test the role of Beclin 1, we overexpressed beclin 1 in IB3-1 cells. HA-tagged beclin induced increased LC3 II and reduced p62 accumulation (Fig. 11A) and increase of autophagosome formation (Fig. 11A); moreover, HA-beclin was detected at the ER level (Fig. 11B, top panel) and restored co-localization of hVps34 (Fig. 11B, bottom panel), hVps15, Ambra1, and Atg14L (data not shown) with calnexin

CHAPTER II

FIGURE 10. Sequestration of beclin 1 interactome in aggresomes drives defective autophagy in CF airway epithelia

CHAPTER II

FIGURE 11. Effect of Beclin 1 overexpression

CHAPTER II

These findings suggest post-transcriptional regulation of beclin 1 expression in CF airways. Since beclin 1 protein sequence contains QP and QXXP motifs, specific target sites for TG2 activity (40), we tested the hypothesis that beclin 1 might undergo TG2-mediated cross-linking and in Beclin 1 immunoprecipitates from IB3-1 cells we observed positive isopeptide immnureactivity that were reduced upon TG2 knock down as well as by the TG2 inhibitor cystamine (40) (Fig.12A) or the calcium chelator BAPTA-AM (24,54). TG2 siRNA or cystamine restored the soluble forms of beclin 1 and beclin 1 interactor proteins (Fig. 12B) and reduced also, beclin 1/HDAC6 co-localization and aggresome sequestration (Fig. 12B). These results indicate that TG2-mediated cross-linking drives defective autophagy in CF epithelial cells. Because increased ROS sustain high TG2 levels via PIASy-mediated TG2 SUMOylation (54), we incubated IB3-1 cells with EUK-134 (54) and we observed reduced beclin 1 cross-linking and aggresome sequestration (data not shown), autophagosome formation (Fig. 12C), and beclin 1 upregulation 4 or knocked down PIASy (Fig. 12C). CFTR knock down in autophagy-competent C38 cells reduced beclin 1 and LC3 II protein, and p62 (Fig. 12D). These effects were prevented by cystamine as well as by EUK-134 (Fig. 12D) or NAC (data not shown)These results indicate that ROS-mediated TG2 SUMOylation is responsible for defective autophagy in CF epithelia.

CHAPTER II

FIGURE 12. TG2-mediated crosslinking of beclin 1 induces aggresome sequestration of beclin 1 interactome and drives defective autophagy in CF airway epithelial cells

CHAPTER II

These results indicate that defective CFTR plays a pivotal role in driving beclin 1 downregulation and defective autophagy in CF airways via ROS-TG2 pathway. Next, we investigated whether defective autophagy might play a role in CF inflammation. HA-tagged-beclin 1 rescued pro-inflammatory phenotype of IB3-1 cells as it reduced p42-44 phosphorylation (Fig. 13A), increased 55 kDa PPAR͎ protein (Fig. 13A), and reduced TNF- / $2BC(C* $1%( &*( D investigated whether defective autophagy might in turn sustain TG2 activation. HA- beclin 1 decreased ROS (p<0.01) and TG2 SUMOylation (data not shown), PIASy and TG2 proteins (Fig. 13A). Since targeting TG2 restores inflammation in human and mice CF airways, we investigated whether these effects were mediated by rescue of autophagy. The effects of cystamine or antioxidants (data not shown) in controlling p42-44 phosphorylation, and TNF- secretion (Fig. 13B) were neutralized upon beclin 1 knock down. This indicates that rescue of beclin 1 and autophagy mediates the effects of cystamine and antioxidant molecules in CF epithelia. These results indicate that targeting ROS-TG2 axis ameliorates human and mouse CF airway phenotype by restoring beclin 1 and autophagy.

FIGURE 13. Restoring beclin 1 and autophagy rescues CF phenotype in IB3-1 cells.

CHAPTER II

1.6.In vivo model

The data described above in CFTR-mutated cell line or normal given a strong concept that in CF induced an activation of ROS-TG2 axis that leads to a pro- inflammatory environment. To demonstrate the importance of this pathway in CF, we used others two biological model: ex vivo human nasal polyp mucose and mice model of CF disease. The first one was the more representative model of human pathology since give a correlation between pathological phenotype and genotype of patient considering the thousand known mutation (divided in six classes). The mice model were the fundamental approaches for therapeutic valence of study. The data of TG2 high protein level and activity on nasal polyp was shown by confocal microscopy in fig.14A, and the effects of inhibition of TG2 on inflammatory target, p42\44 and Tyr phosphorilation, was shown in fig. 14B. The fig.14C shown the PPAR͎ level in control and CF nasal polyp and the effect of R283 (fig 14C). To investigate whether TG2-PIASy interaction and TG2 SUMOylation may take place in human airways of CF patients and whether it was induced by the oxidative stress, we used a well established tissue culture model of biopsies of human CF nasal polyps (72).

FIGURE 14. TG2 and PPAR in “ex vivo” human model 36

CHAPTER II

These data allow to validate this experimental model (24,72) and reported that increased TG2 levels are a feature of CF nasal polyp mucosa and that the inhibition of TG2 is effective in controlling mucosal in\ ! % els of PPARy protein (24). We found that TG2-PIASy colocalized in human CF airways (Fig. 15A) and that this interaction was inhibited upon treatment with the EUK-134 (Fig. 15B). After EUK-134 treatment the distribution of TG2 in CF nasal polyp biopsies was similar to that observed in non-CF controls (24).

FIGURE 15. ROS-mediated PIASy-TG2 interaction in human nasal polyp mucosa.

CHAPTER II

This suggests that the inhibition of TG2-PIASy interaction restores the physiological levels and distribution of TG2 in CF airways (22). Furthermore, in CF nasal polyp mucosa TG2 colocalizes with SUMO-1, and the incubation with EUK-134 inhibited TG2-SUMO-1 colocalization (data not shown). Non-CF control nasal polyp mucosa showed very faint TG2, PIASy, or SUMO-1 expression at the epithelial level (Fig. 15C and D). To test the effects of TG2 inhibition in vivo, we treated CF mutant mice homozygous for F508del-CFTR mutation (73) and their control littermates (36) with cystamine, previously reported to inhibit TG2 and ameliorate disease manifestations in a mouse model of Huntington’s disease (38). We treated CF and wild-type mice with a daily injection of cystamine or PBS, for 7 days. After treatment with PBS the expression and distribution of the tested markers remained unaltered as compared with the pattern observed in untreated CF mice. Before treatment, as well as after treatment with PBS, all seven tested CF mice showed increase of ROS (FIG.16A), as well as TG2 activity (Fig. 16B), as compared with control littermates. PPARy was also reduced and sequestered in aggresomes (Fig. 15B), whereas phosphorylation of p42–44 (Fig. 16B) and increase of TNF-a protein were observed (Fig. 16C). Cystamine did not induce any changes in wild-type mice (data not shown). This cascade of events leads to reduced secretion of inflammatory cytokines as TNF- / 2 oattractant IL-8. As above described, targeting ROS-TG2 axis ameliorates human and mouse CF airway phenotype by restoring beclin 1 and autophagy, therefore we used ex-vivo and mouse model to validate the data of CFTR defective cell line, and control. Defective basal autophagy was also observed in human nasal polyp biopsies from severe CF patients (n=10) and mouse lung tissues from F508del-CFTR homozygous mice (CftrF508del) (73). Reduced beclin 1 protein levels were also observed in CF nasal polyp biopsies (24,54,72) (n=10) and in airway tissues from CftrF508del mice (n=10) (Fig. 17A). To investigate whether rescuing beclin 1 was effective in restoring autophagy in vivo in CF airways, CftrF508del mice (n=3 per group) were administered intranasally with a lentiviral vector encoding beclin 1 (74,75) or a lentiviral vector expressing the GFP. Overexpression of beclin 1 from the lentivirus induced significant increase of LC3 dots (p<0.001), reduced p62 accumulation, TG2 protein level and activity as compared to CftrF508del mice treated with LV-GFP (Fig. 17B). To study the effects of Cystamine or NAC in vivo, we treated CftrF508del mice with daily intraperitoneal injections of NAC (n = 10) or PBS (n = 10). Treated mice showed a pattern similar to cell line.

CHAPTER II

FIGURE 16. TG2 inhibition in mice model

CHAPTER II

FIGURE 17. Restoring autophagy by cystamine or beclin 1 overexpression rescues CF phenotype in mice CF airways

The data obtained reveled a direct link between the regulation of IL8 and inflammation TG2-ROS driven. The IL8 was a drive key in neutrophils movements, and the data obtained about its regulation and modulation in inflammatory environment offer a mechanism for the study of migration IL8-induced and can to develop a model for biotechnological application. For this reason, in second part of activity, we developed a 3D motility model for study of migration driven by chemoattractant molecule, and we employed an important system for screening of potential molecules in therapeuthic application

CHAPTER II

2.Chemotaxis: Motility model

Cell motility is involved in a number of biological processes, from morphogenesis to inflammation. In spite of its relevance in human physiology the currently available techniques for the characterization of cell motility do not allow clear advantage in quantitative interpretation and application in biotechnology for industrial purposes. Furthermore the automation of measurement and a contemporary analysis of multiple samples are controversial . The application of new technologies, as the time-lapse microscopy (available in our lab) has revealed to be useful in overcoming these limits and allowed quantification of the random or directional (chemotaxis) cell movement as well as haptotaxis or contact guidance phenomena (1) and represents the nearest approach to in vivo assays. Cell migration is a very complex phenomena that are strictly related to the inflammatory response as well as several processes of cell biology and development. Chronic inflammation is an ideal condition for the study of cell migration and its deregulation. In this condition several cell populations are recruited in response to the enhanced local release of chemo-attractant molecules (chemokines) that favour the infiltration of circulating leukocytes within a disease-target tissue. This phenomena is known as chemotaxis and the chemical mediators (chemokine, citokine) recall the white cells from blood; the cells must go through different layers (blood vessel, extracellular matrix, epithelia). The formation of chemical gradient (cytokine- chemokine) is an important driving force for cell migration. The first problem in studying the motility is the creation of physiological habitat for cell. The 2D model or microscopy approaches are not a real condition of organisms where the cells interact with other cells (tissue), matrix or liquid (blood cells). For this reason the generation of a 3D model system is mandatory. During inflammation, the neutrophils are recruited in the site of injury by chemoattractant molecules that create a gradient flux in tissue; this gradient induce a extravasion of neutrophils from blood vessel and a movement to site of inflammation. These molecules are the chemokines (interleukin 8, IL-8) which react with specific receptor present on cell membranes to activate their movements. After extravasation, the neutrophils need to cross over a 3D structure, rich in fibronectin, collagen, elastin, the extracellular matrix. This structure is very complex and alters the movement of cell and was subjected to continuous modeling. Many membrane proteins interact with it (i.e. integrin) inducing a rearrangement of cytoskeleton.

2.1.The 3D matrix collagen gel

The first experiments were focused on the study of a ideal 3D model to observe neutrophil motility in basal condition, as well as in pathological condition. The literature offers different models to study motility in 3D such as Boyden assay (76) (a typical indirect assay), that was ideal to perform an initial screening, but the results can be affected by variations in adhesion, motility or spatial orientation. For

CHAPTER II

FIGURE 18. Schematic representation of 3D model: matrix collagen gel and chambers

CHAPTER II example, an increased motility (chemokinesis) would result in an increased migration of randomly directed cells and might score as chemotaxis in many studies. To perform 3D motility experiments, we created a matrix gel to mimic the extracellular matrix and to allow a movement of cell without impediments. To prepare the gel we used collagen rat-tail high concentration, to dilute in different percentage. We added in the mix a precise quantity of medium 1X and 10X free, to allow the cell viability, and we used NaOH 0,1 M to stabilize the pH around physiological value (pH 7.2). All components were kept in ice to prevent gelification of mix during preparation. The gelification required 30 minutes in the incubator at 37°C and 5% CO2. In first analysis we used a multiwells culture dish to test the polymerization of gel. The confocal microscopy images reveled an auto florescence and a fine structure of collagene fibres (Fig.18A). A fibroblast cell line (HT1080) was used to verify cell vitality and the absence of constriction in movement. In second analysis, to study the chemotaxis phenomena, we projected and realized a prototype of chamber to surpass the limits of Boyden assay. It is a rectangular and horizontal chamber splits in two compartments by a divisor barrier. It is composed by two aluminum blocks: A) charatterized by one well for the matrix gel with cells, B) have two wells, one for control gel without the cells and other one for gel with chemoattractant; the two blocks are separated by a semipermeable membrane to allow the flux of chemoattractant. The chamber is equipped with a coverslide on the bottom (for microscopy analysis) and a lid to close the chamber (to keep sterile conditions). The two block are united by some screws and a little quantity of silicon grease, to close all slits (Fig.18B). More experiments underline a objective limit in assembling of chamber. On the basis of these observations, we realized a second prototype of chamber composed by single aluminum block with one single well, in the center; this well is “subsplited” in two wells by a little aluminum block covered with a semi-permeable membrane of 0.22 μm pores (Fig.18C). The assembling processes is more easy. To demonstrate the efficiency of chamber, we performed more experiments of gelification test and cell viability test. We kept the cells in the chambers at 37°C and 5% CO2 for 24h or more, and observed the viability of the cell with TIME LAPSE microscopy images. The Fig.19 showed the HT1080 cell line in different layers of the gel at time 0h and after 24 hours; these images reveled any alterations of cells morphology and composition of matrix gel.

CHAPTER II

FIGURE 19. HT1080 in collagen gel

2.2. TIME-LAPSE system

To study the 2D cell motility phenomena ,fluorescence microscopy techniques was available on fixed samples, where the examined cells are observed in instants of time that precede and follow the addition of particularly factor, missing the progression of motility event. The advent of technology as time-lapse microscopy has allowed to overcome these limits, being able to follow cell motility from time to time and obtain quantification of the random or directional (chemotaxis) cell movement. Microscopy images give a static moment of processes lacking a intermediate time of phenomena. Therefore to study complex phenomena such the chemotaxis or cell with high motility a system able to follow all time of neutrophils\ cells movements was required, as the time lapse video system. The TIME LAPSE system was based on an inverted optical microscope, equipped with mercury lamp and epifluorescence reflectors. The microscope was placed on an anti-vibrating table and was equipped with motorized stage and focus, that allowed to automatically position the field of view within the sample under observation. Images are acquired using a monochrome CCD camera and sent to the PC through Firewire interface. In order to mimic environmental conditions required to run time-lapse experiments, a microscope incubator was used, consisting in a plexiglass box, coated with a

CHAPTER II

FIGURE 20. TIME-LAPSE work station

polystirol thermal isolator, in the box a radiator and a fan allow heat transfer and temperature uniformity. The temperature in the incubator was monitored by a thermocouple, connected to a PC via serial port. Temperature control was ensured by a software PID control that continuously update the temperature of the radiator. A 5% CO2 air flow is humidified through a bubbling column inside the incubator, and sent in a micro chamber that keeps the sample in a controlled air, constant temperature environment The system was connected to a video-camera with high solution and velocity to follow the motility of high speed cells, and this was manage by a software on PC. The software allowed to move the table (the sample) on three axis (x,y,z) to follow the sample in all 3 dimension. The x and y axis was controlled by a joystick connect to motor of table, and the z axis was controlled by a system of focus. Using the software we selected a specific object with a defined x, y position and chose a

CHAPTER II position in the gel that corresponded to a layer with cell on focus; with specific functions of software we selected the total delta of layer (in μm), the number of layers that const the total delta and the time of acquisition: a Z-stack acquisition. In this way we followed the single cell (of selected object) in 3D for every single instant of time(Fig.20). Time lapse experiments ran according to the following protocol. Once introduced the samples in the incubators, image acquisition started. A software allowed periodic scanning of marked positions in the sample. Input parameters was the sample positions to image and the acquisition time frequency. Each time cycle the software was then control the microscope stage and focus in order to cyclically place any marked position under the field of view and automatically acquire images, that was stored on hard drive, the software was than wait a pause up to the next acquisition cycle. At the end of the experiment we obtained a number of images for every single layer in single instant of time, and we could create a sequences of images as like a film. The single cells present in the images were analyzed by a specific software of cell analysis (Image Pro-Plus) provided of a macro to follow the single cell; through a mathematical model we obtained many parameter such as speed, dimension, trajectory, diffusion profile (in presence of chemoattractant molecules).

2.3.Calibration of system

The diffusion profile of chemoattractant was basic to analyze the cell motility in 3D model. The concentration gradient was studied by measurement of concentration of a fluorescence type of chemoattractant, IL8, or by a component with analogous molecular weight that take bound the fluorescent tag (dextran bound to FITCH). The measurement of fluorescence was perform by means of optic microscopy and images analysis by using the Imege ProPlus software. A preliminary calibration was performed on sample with different concentration of FITCH-dextran without the barrier, in order to have a uniform distribution. The sample exposition at excitant radiation was minimize to avoid the photobleaching of tag. In second time we performed the experiments by using different concentrations of FITCH-dextran. We kept in the well a mix collagen with FITC-dextran and observed the diffusion trough the membrane in control well with gel alone. The emitted fluorescence was proportionally direct to concentration of FITCH-dextran used (Fig.21A). The intensity of fluorescence was expressed as mean of grey level. We calculated the concentration profile of FITCH-dextran during time and in the space, by measuring the emitted intensity of sample light into control well at different distance from membrane, in chancing of time since the adding of FITCH-dextran. The fitting data revealed a diffusion coefficient of 2·10-6 cm2/sec, according to literature data (Fig.21B).

CHAPTER II

FIGURE 21 Diffusion of FITC-dextran: calibration system

CHAPTER II

2.4. Neutrophils in collagen gel: 3D model

To test whether the chamber was an ideal condition to study neutrophils as well as other cell line (epithelial, fibroblast, gut, etc.) we performed a experiment with neutrophils isolated from peripheral blood. The neutrophils were isolated from blood followed a specific protocol with Ficoll- Paque buffer. Ficoll-Paque was normally placed at the bottom of a conical tube, and blood was then slowly layered above Ficoll-Paque. After being centrifuged, the following layers will be visible in the conical tube, from top to bottom: plasma and other constituents, a layer of mono-nuclear cells called buffy coat (PBMC/MNC), Ficoll-Paque, and erythrocytes & granulocytes which should be present in pellet form. This separation allows easy harvest of PBMC's. After washing and Dextran-NaCl lysis, the nutrophils were isolated; an appropriate number of white blood cells(5*105 cell\ml) was counted in Burcker chambers. The mix of collagen was prepared in a conical tube without any cheattractant, and was divided in two mix: one of them was put into control well and other one was mixed with appropriate number of cell. In this way the cells were included in collagen fibres reticulum that block the sedimentation and represent a tridimensional substratum and can partially reproduce the microstructure of extracellular matrix. After gelification at 37°C and 5% CO2, the chambers was fixed on specific supporter of TIME LAPSE STATION. The cell motility was observed for 24h with Z-stack acquisition images every five minutes (Fig.22A). The quantitative analysis of single cell performed by Image Pro-Plus software, reveled a random motility in absence of chemokin (Fig.22B). Moreover we performed the experiments in same conditions (number cells, acquisition time) in presence of IL8 (50μg\ml); its diffusion time was very early and neutrophils showed a directional motility. The results were obtained from mathematical analysis of motility parameters. The qualitative and quantitative data analysis evidenced a directional motility of neutrophils under IL8 diffusion compared to absence of IL8. The results indicated that the prototype chamber was an ideal model to study migration of different cell lines. The qualitative and quantitative analysis of images allow to obtain many information about the “parameters” motility-dependent. Moreover, was more hopeful for application in industrial biotechnological sector. Furthermore, the system offered a new approaches to study the motility IL8-induced and IL8 modulation (see above data) represented a key strategy for new therapeutic application.

CHAPTER II

FIGURE 22. Neutrophils in collagen gel

CHAPTER II

2.5.T84 motility: 2D model

The 3D model was one of many way to study cell movements in a physiological system with a collagen gel to mimic extracellular matrix. the neutrophils migration under chemokine flux offered a valid model to set the system condition, and modulation of chemoattractant factors was the key to study chemotaxis phenomena and develop biotechnological application. CF disease represented the optimal system to study the mechanisms of modulation since it was characterized by a strong recruitment of neutrophil under IL8 diffusive flux CFTR-defective driven. The inflammatory environment was crucial effector of this condition such as for many other patholy as well as Coeliac Disease (CD). The common characteristic of either CF and CD, besides the inflammation condition, was the presence of high TG2 level. In CD the inflammatory milieu was a response to a persistence attach from its inductor, the gluten, which led to an evident modification of tissue (or cell line model) architecture with consequent re-arrangement of cytoskeleton. This data offered a valid condition to develop a 2D-motility model to study epithelial cell line in adhesion and view the cytoskeleton modification. Coeliac disease is a relatively common pathology that affects 1% of the population, and is precipitated by the ingestion of gluten proteins. In predisposed individuals, gliadin acts as leading to a powerful activation of gliadin speci[ (> fragments of gliadin have been found to be ‘toxic’ for predisposed individuals and it has also been suggested that different portion(s) of gliadin, ostensibly not those recognised by Tcells, modulates this innate activation of coeliac small intestine. Further studies have indicated that TG2 is the essential “pivot” in the pathogenesis of celiac disease (27) because TG2 operates a crucial posttranslational modification,

CHAPTER II

FIGURE 23. Internalization and pro-inflammatory effects of p31-43 and control peptide

CHAPTER II the deamidation of strategic glutamine to glutamic acid, within the sequence of immunodominant gliadin epitopes(Subunits of the antigenic determinant that are most easily recognised by the adaptive immune system and thus most effective in the induction of antibody and T-cell responses). Among toxic peptides , p31-43 induces the expression of early markers of epithelial activation and leads to enterocyte apoptosis (77) both in coeliac intestine and in sensitive intestinal epithelial cell lines. Moreover, p31-43 induces early tissue transglutaminase (TG2) upregulation compared to immunodominant gliadin peptides (35). This indicates that TG2 is not only central to the adaptive response to gliadin as the main deamidating enzyme of the immunodominant epitopes, but is also a key player in the initiation of inflammation in coeliac disease. However, it is still unknown how p31e43 induces early TG2 activation in both coeliac intestine and ‘gliadin- sensitive’ intestinal epithelial cells (35,79). We have demonstrated that in cystic fibrosis (CF) airway epithelia (24) sustained TG2 activation is driven by the increased levels of reactive oxygen species (ROS) caused by the CFTR genetic defect. We therefore investigated whether p31-43 was able to drive an intracellular prooxidative environment and demonstrate that p31-43 leads to an increase of ROS levels in celiac duodenum as well as T84 intestinal epithelial cells. To understand how p31-43 challenge induces oxidative stress, we analysed the delivery of this peptide to the intracellular degradation systems. Recent data demonstrated increased transepithelial translocation of the 33-mer (p56–89) in active coeliac disease compared to healthy controls. Moreover, incomplete degradation of the 33-mer and protected transport of the peptide 31-49 occurs in patients with untreated coeliac disease, thus favouring their respective immunostimulatory and toxic effects (79). Here we show that p31-43 is retained within the lysosomes as compared to the more rapid disappearance of the immunodominant peptides pa-2 or pa-9 from the intracellular compartments. T84 cells were incubated with biotinylated p31-43, p-a2, p-a9 for 15, 30, 60, 90 min, 3 h and 24 h at 378C. Soon after 15 min of challenge, peptides co-localised with EEA-1 (Fig. 23A), a marker of early endosomes; After 90 min the internalized peptides co localised with Lysosomal Associated Membrane Protein-1 (LAMP-1) (Fig. 23A), a marker of lysosomes. No internalisation of peptides or colocalisation with LAMP-1 were observed when T84 cells were pre-incubated with M-!-CD, which inhibits clathrin-mediated endocytosis (80), or filipin, a specific inhibitor of lipid raft- or caveolae-dependent endocytosis (80) (data not shown). This suggests that endocytosis is essential for the delivery of gliadin peptides to the lysosomes for degradation. Increasing appearance of the internalised p31-43 in LAMP-1- positive vesicles, mainly limited to perinucelar localisation, was observed after 3e24 h of challenge (Fig. 23A). At these late time points p-a9 (Fig. 23A) or pa-2 (data not shown) were only faintly detected in LAMP-1-positive structures, thus indicating that p31e43, but not p-a9 or p-a2 is retained within the lysosomes. Moreover, after 90 min of challenge neither p31e43 nor pa-9, induced significant increase of ROS (Fig. 23B). At the later time points (3 e 24 h of challenge), p31-43, but not pa-9 (Fig. 23B), induced a time-dependent significant increase of ROS levels, suggesting that the prolonged persistence of p31-43 in LAMP-1 positive vesicles generates a pro- oxidative environment. Blocking p31-43 endocytosis with M-!-CD, prevents the increase of ROS (data not shown). The generation of a pro-oxidative environment induces activation of different stress sensitive signalling pathways (78,24); therefore, we investigated whether p31e-43- induced ROS-mediated p42/44 MAPK phosphorylation in T84 cells. We

CHAPTER II demonstrated that p42\44 MAPK phosphorylation induced by the challenge with p31e43 was prevented by the incubation with the catalase-superoxide dismutase (SOD) mimetic EUK-134 (Fig.23C). To investigate whether the increased levels of ROS may account for the p31-43-induced increase of TG2 levels,(24)we incubated p31-43-challenged T84 cells with EUK-134. We demonstrated that EUK-134 was effective in controlling the increase of TG2 protein levels (Fig.23C) and TG2 activity (data not shown) induced by p31e43. No increase of TG2 was observed after challenge with p31-43 upon M-b-CD or filipin treatment (data not shown). This indicates that p31-43 internalisation is critical for the induction of ROS-mediated TG2 activation. The pro-inflammatory environment p31-43-driven induces a strong re-organization of cytoskeleton; as matter of fact the lumen intestinal of CD patients shown atrophy of villi and cryptic hyperplasy due to persistence T-cells infiltration gluten-toxicity induced. To test whether the p31-43 toxicity could impact on cell motility and interaction of gut epithelial cell line (T84), we used a 2D model. The T84 cell line was characterized by a typical collective migration where the cell line was organized in a compact structure such as “island”, and their movement was due to movement of total and contemporary motility of single cell that composed the island. During the movements, the cells\islands have interaction either between single cell, either between the island, and we showed a phenomena like “englobetion” of island after their touching. To demonstrate the role of p31-43 on cell motiliy, we cultured the T84 cell line on multiwells cultur dish; we incubated p31-43-challenged T84 cells with EUK-134 or with TG2 monoclonal antibody, 4G3, to block the toxic effect with specific inhibition of ROS or TG2 We performed a 2D motility experiments using time-lapse workstation and setting the acquisition time every 15 minutes for a total of 48 hours. The images reveled a visible motility alteration ,evidenced by a strong collective migration of cell and faster “englobetion” of island in presence of p31-43versus the control not treated; moreover, in presence of inhibition the motility rescued its pattern like not treated.(Fig.24A) To demonstrate whether the alteration of motility p31-43 induced, was due to reorganization of actin, we fixed the samples and used a specific fluorescent molecule, FITCH-phalloidin, to marker the cytoskeleton. The confocal imges reveled a tipical actin structure of T84 in control not treated and a rearrangement of cytoskeleton after p31-43; the actin rescued the structure after inhibition of ROS or TG2.(Fig.24B) The results indicated that TG2-ROS axis was critical point to regulation in Coelic Disease, and its modulation will be the goal to treatments of pathology; moreover, the 2D model analysis open the way to possible system of screening of molecule with high impact on motility and become a future fast test screening in biotechnological application.

CHAPTER II

FIGURE 24. T84 motility and cytoskeleton organization: 2D model

CHAPTER III

CONCLUSIONS

The results of PhD thesis project give an important contribution in defining the pathological mechanisms of airway inflammation in CF and identifying new therapeutic options for patients . The principal results of the study have been published in four papers on international journals (J. Immuol 2008, J. Immunol 2009; Gut 2010; Nature Cell. Biol. 2010) and underline: A) the role of TG2 as a new pathogenic factor in Cystic Fibrosis (24). The novel [ / % of the study is the demonstration that CFTR mutations lead to an increase of TG2 levels and activity that, in turn, down-regulate the anti-in\ammatory effects of PPARy. We have shown that PPARy is not only reduced in CF epithelium and CFTR-defective cell lines, but also sequestered in perinuclear aggresomes, prominent pathological features common to many neurodegenerative conditions such as Huntington and Parkinson’s diseases (81). Aggresomes are generated in response to alterations of the ubiquitin proteasome system, the principal cellular mechanism of degradation of misfolded proteins and represent a cytoprotective mechanism that cells adopt to clear mutant/modified aggregate-prone proteins (82). This suggests that the formation of the aggresomes in CF is caused by the accumulation of misfolded proteins fated to [ /% / ( TG2 is a pleiotropic enzyme with a calcium-dependent transamidating activity that results in cross- - %2 (y-glutamyl) lysine bonds (36,40). Our results demonstrated a signi[ 2 / 51 epithelium and CFTR-defective cell lines. TG2 up-regulation was also responsible for the aggresome formation in CF. Importantly, TG2 induced cross-linking of PPAR͎ and its sequestration into aggresomes, because the TG2-inhibitor R283 produced a signi[ / EE&"͎ protein aggregates and restored a wider intracellular protein distribution. Collectively, our study provides a molecular explanation of the reported association between CFTR defective function and sterile inflammation. Highlighting the role of TG2 in CF, we have also indicated novel ways to modulate inflammation, a key component in CF pathogenesis.

B) the role of post-translational modifications of TG2 as a link between genetic defect of CFTR ad inflammation (54) These data have underpinned the relationship between CFTR defective function, oxidative stress, and chronic airway inflammation in CF. We have identified TG2 SUMOylation as a pivot in driving the pro-inflammatory CF phenotype. SUMOylation has been defined as a key player of the posttranslational network to regulate key cellular functions. Herein we demonstrate that in CF airway epithelia the increased levels of ROS lead to TG2 SUMOylation via interaction of TG2 with PIASy, an E3 ligase. Oxidative stress increases PIASy protein levels and favors TG2 SUMOylation that leads to the persistence of high TG2 tissue levels by down-regulating TG2 ubiquitination and proteasome degradation. Our results also indicate that TG2 SUMOylation is a general response to oxidative stress. The rheostat role of TG2 makes this enzyme an attractive target to restore cellular homeostasis and dampen chronic inflammation in CF airways. The regulation of the high levels of TG2 protein or the inhibition of sustained TG2 enzyme activation may represent a new attractive approach to control disease evolution in CF patients. To

CHAPTER III evaluate whether we can translate our in vitro [ / % 222 model and con[ !% vance in vivo we have studied CF mutant mice homozygous for F508del-CFTR. Our results highlight TG2 as an unforeseen unifying link between genetic defect, deregulation of cellular homeostasis, and in\ammation. They also indicate TG2 as a candidate target for the design of a pathogenic-based therapy in CF and add to the rationale for attempting inhibition of TG2 in CF patients..

C) the mechanisms of autophagy inhibition via ROS-TG2 axis , in epithelial airway cell line and mice model of CF. For the first time this study defined the CF as an autophagy-related disease (87). Defective autophagy is a critical mechanism in several chronic human diseases (83) such as neurodegeneration (84) and cancer. Here we show that defective autophagy due to decreased levels of beclin 1 protein, a key molecule involved in autophagosome formation (60,63,66,70), drives lung inflammation in CF. A common feature between this neurodegeneative disease and CF is the aberrant intracellular accumulation of misfolded proteins within the affected tissues (85,86). In CF airways, misfolded or damaged proteins, such as misfolded CFTR and PPAR͎, accumulate into aggresomes. Here we demonstrate that in CF epithelia TG2- mediated crosslinking and sequestration of beclin 1 dislodges PI3K platform from the ER, thus inhibiting initiation of autophagy. TG2-driven defective autophagy in turn increases ROS and TG2 levels, thus generating a vicious cycle leading to a deleterious pro-oxidative and pro-infammatory environment. Autophagy inhibition with p62 accumulation (58) increases the levels of proteasome substrates and compromises the Ubiquitin Proteasome System (UPS) (58), thus favouring aggresome formation. TG2 contributes to proteasome overload and aggresome formation in CF airways by inducing cross-linking and aggregation of several substrate proteins. Therefore, TG2 functions as a rheostat of the post-translational network and UPS under disease conditions and switch off the post-translational regulatory mechanisms. As a consequence of this function, TG2 is involved in the pathogenesis of neurodegenerative diseases due to protein aggregates. We have previously reported that CFTR inhibition results in the upregulation of ROS and their downstream events leading to inflammation. Here we show that CFTR knock down drives defective autophagy as it occurs in human and mouse epithelia carrying either homozygotes for F508del or compound heterozygotes for two severe CFTR mutations. Our results provide a new mechanism linking CFTR defect and inflammation via ROS-TG2-mediated inhibition of autophagy. Therefore, we have provided a novel rationale and the mechanism of action for both NAC and cystamine in treatment of CF patients. In conclusion, our study suggests that restoration of beclin 1 and autophagy may be a novel approach to treat CF and may pave the way for the development of a new class of drugs that by enhancing beclin 1 levels could be effective treatments for CF.

D) the development of model for 3D migration assay and 2D motility The study of mechanisms that regulate the IL8 is mandatory to develop a 3D model for IL8-induced neutrophils migration assay for potential biotechnological application. The 3D model is a ideal model to study the movements of cell in a 3D matrix of gel where a diffusive flux of molecule induced migration; instead, a useful model to test the cell motility in adhesion condition was the 2D model. The 2D model is ideal condition to study effect of molecule that influenced cell motility due to a organization of cytoskeleton; for this propose, the intestinal epithelial cell (T84), model of Coeliac

CHAPTER III

Disease, represent an ideal application of 2D the model. The Coeliac Disease is characterized by sensitive response to A-gliadin peptide fragments, that induce an immune activation with an inflammatory response. Among the gliadin peptide, the p31-43 induces an alteration of inflammation via ROS-TG2 pathway (82), which have many implications in the organization of actins filaments. This pathology is strong linked to TG2 that has a crucial role in the pathogenesis of disease. The 2D model offers a ideal screening test to verify whether the p31-43 toxicity can impact on cell motility and interaction of epithelial cell line (T84). The results highlight the possibility to use this model as a prototype of fast screening procedures to test the effects of several putative toxic fragments as well as molecules that prevent gluten-induced toxicity.

These data provide the rationale for the development of cell motility and migration assays as ideal instruments to test the efficiency of molecules with potential pharmacological impact. The use of quantitative and qualitative analysis could also be implemented in the screening of various drugs potentially effective in controlling cancer cell invasiveness and in modulating drugs-resistance. The effects of TG2 and ROS modulation on these dynamic processes could allow the set up of appropriate devices for either research purposes or industrial biotechnological application.

CHAPTER IV

MATERIALS AND METHODS

Cell lines and cultures. Human CF bronchial epithelial cell line IB3-1, carrying F508del/W1282X CFTR mutation and isogenic stably rescued C38 (LGC Promochem, Milan, Italy), were cultured with LHC-8 medium (Invitrogen) supplemented with 5% FBS and the appropriate amount of penicillin/streptomycin. Normal bronchial epithelial 16HBE (LGC Promochem, Milan, Italy) were cultured with MEM Earl's salt medium supplemented with 10% FBS and 1% penicillin/streptomycin, as recommended by American Type Culture Collection. CFBE41o- human CF bronchial epithelial cells (gift from Dr. D. C. Gruenert) were cultured MEM Earl's salt L-Glutamine (200 mM L-Glutamine) medium supplemented with 10% FBS and the appropriate amount of penicillin/streptomycin. Cell lines were incubated for 24 hours with R283 (+C)?), μ $+C)?>% &/ *0- acetyl Cysteine (NAC, 10 mM, Vinci Biochem), EUK-3$+C)%/ml, Vinci Biochem), rapamycin (100 nM, Sigma Aldrich), chloroquine (25μM, Sigma Aldrich), E64d (50 )%/ml for 4 ho * /22 &$+C)%4 3 *"% $C)?3 hours, Vinci-Biochem); BAPTA-&?$+)%/ml, Calbiochem, for 4 hours), proteasome inhibitor MG132 (+C )? 5 ! G *. C38 and 16HBE cells were incubated for 48 hours with CFTRinh172 (20 μM, Calbiochem) followed by cystamine or EUK-134. Human colon adenocarcinoma T84 were cultured in DMEM:F12 mixture 1:1 medium supplemented with 5% FBS and the appropriate amount of penicillin/streptomycin, as recommended by American Type Culture Collection (ATCC). The cells were challenged with p31-43 for 24 h in the presence or absence of the ROS scavenger EUK-134 (50 mg/ml; Alexis Biochemicals), or the TG inhibitor cystamine (400 mmol/l; Sigma). To study internalisation of peptides, cells were challenged for 15, 30, 60 or 90 min, 3 or 24 h at 378C with the biotinylated peptides (20 mg/ml). To inhibit endocytosis, the cells were pre-treated with the cholesterol-binding agent methyl-bcyclo-dextrin (M-b-CD, 10 mmol/l; Sigma) or [lipin (5 mg/ml; Sigma) and then challenged with biotinylated peptides

Human subjects and ex vivo cultures of nasal polyp mucosal biopsies. Ten CF patients carrying severe CFTR mutations (Table 1) and ten consecutive non-CF patients (mean age 21 years, range 16–32) underwent surgical treatment for nonallergic nasal polyposis were enrolled. Nasal polyp biopsies from seven patients for each group (patients #4-10 of Table 1) were cultured for 4 hours with or without EUK- 3 $+C )%4* 0&5 $C ?* $3CC ?* "# $+C3?*( 0 polyp biopsies from five CF patients (patients #6-10 of Table 1) were also incubated with NAC or Cystamine or PPARyantagonist GW9662 (1 μM; Alexis Biochemical), or R283 for 24 h, followed by GW9662 (1 μM) for 4 h. Informed consent was obtained from all subjects, and the ethical committee of Regione Campania Health Authority approved the study.

CHAPTER IV

Mice and treatments. Young adult female CF mice homozygous for the F508del mutation (abbreviated CftrF508del) in the 129/FVB outbred background (obtained from Bob Scholte, Erasmus Medical Center Rotterdam, The Netherlands), and their wild-type littermates were housed in static isolator cages at the animal care specific pathogen free facility of Charles River (Varese, Italy). To prevent intestinal obstruction CF mice were weaned to a liquid diet (Peptamen; Nestle´ Nutrition). Peptamen was replaced daily. The genotype of each animal was checked at 21 days of age. These studies and procedures were approved by the local Ethics Committee for Animal Welfare and conformed to the European Community regulations for animal use in research (CEE no. 86/609).We treated CftrF508del homozygous mice by intraperitone I $(2(*; / /CC)C(C? E> solution for 7 days16 (n=7 for each group of treatment) or NAC with a daily dose of 100 mg per Kg (body weight) in PBS solution for 5 days or with PBS solution (n=10 for each group of treatment). Mice were then killed by i.p. injection of 20 mg of sodium pentobarbital (Abbott Laboratories). Airway administrations of lentiviral vectors expressing the mouse beclin 1 or GFP were performed in anesthetized CftrF508del homozygous mice aged 4-8 weeks as previously described (74). Mice were sacrificed 5 days post-vector administration and lung tissues were collected for analyses.

Plasmids and transfection. The pcDNA3-HA-beclin 1 expression vector (gift from Dr. N. Mizushima) and the corresponding empty expression vectors were used for transfection experiments. IB3 cells were transfected with pcDNA3-HA-beclin 1 and empty vector , as control, using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s protocol. Briefly, one day before transfection, we plated cells in growth medium without antibiotics such that they will be 60-70% confluent at the time of transfection; the day of transfection we prepared the complex of DNA and Lipofectamine 2000 (in a ratio of 1μg:3μl, DNA:Lipofectamine) and incubate for 30 minutes at Room Temperature; we added the complex to the cells and incubated at 37°C for 24 h.

RNA interference. Cells were transfected with 50 nM human TG2, SUMO-1,PIASy, beclin 1 and CFTR siRNA (Invitrogen) or scrambled oligonucleotides by using Lipo RnaiMax (Invitrogen) according to the manufacturer’s instructions. Briefly, one day before transfection, we plated cells in growth medium without antibiotics such that they will be 30-50% confluent at the time of transfection; the day of transfection we prepared the complex of siRNA oligonucletides and LipoRNAiMAX and incubate for 30 minutes at Room Temperature; we added the complex to the cells and incubated at 37°C for 72 h. For siRNA oligonucleotides sequences see Supplementary Information, Table 2 .

CHAPTER IV

Adenoviral Vector. Human MnSOD cDNA was cloned into the shuttle vector pAd5CMVK-NpA (gift from Dr. Michael Brownlee). Briefly, one day before infection we plated cells in growth medium without antibiotics; the next day, the cells were infected with MnSOD or control adenovirus for 2 hours at 37°C.

Quantitative RT-PCR. Total RNA was extracted using the RNeasy Mini Kit (Qiagen). The mRNA was reverse transcribed by SuperScriptTM III First Strand Synthesis System (Invitrogen). Quantitative RT-PCR was performed using iCycler iQ Multicolour Real-Time PCR Detector (Bio-Rad) with iQ TM SYBR Green supermix (Bio-" /*(982 % ; /!-actin levels in the same sample. The sequence of primers is reported in Table 3. The relative amounts of mRNA were calculated by using the comparative CT method.

Western blot analysisCells were washed 2 times with ice PBS and then lysed in buffer containg 50 mM Tris- HCl (pH 7.5), 150 mM NaCl, 1% NP-40, 5 mM EDTA, 1 ? 2 / +C ? 0 1 C )%4 22 / C )%4 apoprotein supplemented with protease inhibitors (Sigma), kept in ice for 40 minutes and centrifuged at 14,000rpm at 4°C for 20 minutes. The amounts of proteins were determined by a Bio-Rad protein assay to ensure equal protein loading before Western blot analysis. Fifty micrograms of cell lysate of each sample was electrophoresed through 10% or 8% polyacrylamide gels (BioRad), transferred onto blotting membranes (Polyscreen PVDF, NEN) and immunobloted for specific primary antibody in agitation at 4°C for 12-18 hours. The primary Abs were counterstained by a HRP-conjugated anti-IgG Ab (Amersham Biosciences) for 60 min at room temperature. Proteins were visualized by chemiluminescence (ECL Plus; Amersham Biosciences) and exposed to X-OMAT [lm (Eastman Kodak). The antibodies used for immunoblot analysis are reported in Table 4. The densitometric analysis was performed by Image J software and each data point is expressed as the mean ± s.d. of three independent experiments.

Immunoprecipitation. Cells were washed 2 times with ice PBS and then lysed in buffer containg 50 mM Tris- HCl (pH 7.5), 150 mM NaCl, 0,5% NP-40, 5 mM EDTA, 1 ? 2 / +C ? 0 1 C )%4 22 / C )%4 apoprotein supplemented with protease inhibitors (Sigma), kept in ice for 40 minutes and centrifuged at 14,000rpm at 4°C for 20 minutes. Following cell lysis , 500 μg of whole cell-lysates were incubated in agitation at 4°C for 8-12 h with antibody specific primary antibody. After the addition of Protein A/G-agarose beads, the incubation was continued in agitation at 4°C for 1 h. After washing, the immunoprecipitated proteins were electrophoresed through 8% polyacrylamide gels (BioRad), transferred onto blotting membranes (Polyscreen PVDF, NEN) and analyzed.

CHAPTER IV

Soluble and insoluble fraction. Cells were lysed in buffer containg 50 mM Tris- HCl (pH 7.5), 150 mM NaCl, 0.5% NP-40, 5 mM EDTA, 1 mM phenylmethylsulfonyl /+C?0 1C)%- 22 /C)%-1 apoprotein supplemented with protease inhibitors (Sigma) and centrifuged at 14,000rpm at 4°C for 20 minutes. After centrifugation the soluble (supernatant) and insoluble (pellet) fraction were used in Western blot analysis using anti-Beclin1. The NP-40 insoluble pellet was dissolved in sample buffer 5X, boiled at 95°C for 5’ and resolved onto 10% polyacrylammide gel.

Cell fractionation. IB3-1 cells were collected in cold buffer A (20 mM Tris-HCl (pH 7.4), 2 mM EDTA, 20 mM 2-ME, 1X PMSF, 1 {g/ml inhibitor protease cocktail), homogenized in Potter-Elvehjem pestle and glass tube (Sigma-Aldrich), and centrifuged at 2000 rpm for 15 min at 4°C to obtain nuclear pellets. Supernatants were collected as cytoplasmic fractions. Nuclear pellets were washed with buffer A and resuspended in buffer B (2 mM Na3VO4, 400 mM NaCl, 1 mM MgCl2, 1 mM EGTA, HEPES 10 mM (pH 7.9), 1 mM DTT, 1X PMSF, 1 μg/ml inhibitor protease cocktail) and incubated on ice for 50 min with occasional mixing to extract nuclear proteins. Nuclear extracts were cleared by centrifugation (7000 rpm, 15 min, 4°C), and supernatants were collected as nuclear fraction. Then, cytoplasmic and nuclear whole-cell fractions were analyzed by western blot analysis

Confocal microscopy. Cell lines. Treated or untreated cells were [xed in ice methanol for five minute at -20°C, washed 3 times with PBS, permeabilized with 0.5% Triton X-100 for ten minutes at room temprature, washed 3 times with PBS and incubated with primary antibodies over night at 4°C. Human tissue sections. Five-micrometer frozen human lung tissue sections were [xed in acetone for 10 min. The sections were incubated for 2 h at room temperature with primary Abs. Mice lung tissue. Seven-micrometer frozen lung tissue sections from each mice were [xed in acetone for 10 min. The sections were incubated for 2 h at room temperature with primary Abs. The primary antibodies were counterstained by a Alexa fluor-conjugated anti-IgG Ab. and an LSM510 Zeiss confocal laser-scanning unit (Carl Zeiss, Germany) was used for detection. The antibodies used for immunofluorescence analysis are reported in Table 4.

ROS detection. Cell lines were pulsed with 10 μM 5(and 6)-chloromethyl-2'7'- dichlorodihydro\uorescein diacetate acetyl ester (CM-H2DCFDA; Molecular Probes, Invitrogen), according to manufacturer suggestions, and analyzed with Wallac 1420 multilabel counter (PerkinElmer). Tissue sections (5 μm) of nasal mucosa and cells were incubated with 10 μM CMH2DCFDA, according to manufacturer suggestions, and an LSM510 Zeiss confocal laser-scanning unit (Carl Zeiss, Germany) was used for detection. 61

CHAPTER IV

In situ detection of TG2 activity and protein. TG2 protein and enzyme activity were detected and analyzed by confocal microscopy. Brie\y, for detection of TG2 enzyme activity, live cells were preincubated with TG2 assay buffer (965 μl of 100 mM Tris/HCl (pH 7.4), 25 μl of 200 mM CaCl2) for 15 min and then with the same TG2 assay buffer added with 10 μl of 10 mM biotinylated monodansylcadaverine (bio-MDC; Molecular Probes) for 1hat room temperature. The reaction was stopped with 25 mM EDTA for 5 min. The cells were then [xed in 4% paraformaldehyde for 10 min. The incorporation of labeled substrate was visualized by incubation with PE- conjugated streptavidin (Dako-Cytomation; 1:50) for 30 min. Control experiments included the omission of bio-MDC, as well as replacement of 200 mM CaCl2 with 200 mM EDTA. Blocking experiments were also conducted by incubating cells with the active site inhibitor R283 (250 μM) for 1 h before incubation with the substrates (bio- MDC or biotinylated peptides) and detection of TG2 enzymatic activity. Data were analyzed under \uorescence examination by LSM510 Zeiss confocal laser-scanning unit (Carl Zeiss). The simultaneous detection of TG2 protein and activity was performed by incubating live cells with bio-MDC for1hat room temperature, according to the above described procedure, and then with anti-TG2 CUB 7402 mAb (NeoMarkers; 1:50) for 1hat room temperature. This was followed by simultaneous incubation with PE-conjugated streptavidin (DakoCytomation; 1:50) and FITC-conjugated rabbit anti-mouse Ig F(ab')2 (DakoCytomation; 1:20) for 1 h. An LSM510 Zeiss confocal laser-scanning unit (Carl Zeiss, Germany) was used for detection. Of TG2 activity and protein on tissue sections

Electron Microscopy. Cell culture monolayers were fixed for 15 min at 4 °C with 4% paraformaldehyde and 2.5% glutaraldehyde in 125 mM cacodylate buffer and stained with uranil and lead citrate and examined in a Leo912 electron microscope. We used Image J software (US National Institutes of Health) to quantify the number of autophagosomes per cell in 15–20 different randomly taken micrographs for each condition

ELISA .Human or murine TNF-a secretion was measured using the BD OptEIA TNF- a ELISA kit II (BD Biosciences). Measurements were performed at least in triplicate. 6 Values were normalized to 10 cells; results were expressed as means ± SEM.

Statistical analysis. Data are reported as arithmetic mean ±s.d.. Data distribution was analyzed for normality and statistical analysis performed using the one-way ANOVA. Significant differences are indicated in the figures. All data were obtained from independent measurements. Data were analyzed using SPSS 13 software. Statistical significance was defined as P value of <0.05.

CHAPTER IV

Collagen gel mix. Liquid collagen rat-tail high concentration (9 mg/ml) was diluted to obtain a final concentration of 2mg/ml different percentage. We added in the mix an appropriate amount of medium 1X free and 10X free and we used NaOH 0,1 M to stabilize the pH around physiological value (pH 7.2). The was placed in incubator at 5 37°C and 5% CO2 for 30 minutes. We added 2*10 cell/ml of gel, or FITC-dextran (50μg/ml).

Neutrophils isolation. Peripheral blood (10 ml) was taken from healty human donors into BD Vacutainers containing K3EDTA. Neutrophils were freshly isolated by dextran sedimentation and centrifugation on Ficoll-Hypaque. The pellet of a density- gradient centrifugation containing neutrophil granulocytes and erythrocytes was diluted 1:1.3 with a high-molecular-weight dextran solution. After 2h erythrocytes had settled down and the neutrophil granulocytes containing supernatant was separated from the pellet. If necessary, red blood cells in the neutrophil-rich fraction were lysed with hypotonic saline. The neutrophils were washed twice with phosphate buffered saline (PBS) and then resuspended in RPMI 1640 containing 10% heat-treated fetal bovin serum (FBS). The obtained purified neutrophil granulocytes were used immediately after isolation.

Image analysis. The Image Pro Plus software, with a home-made macro tool, was used to analyze single cell motility. Math-lab software was used to fitting data.

Phalloidin staining. Treated or untreated cells were [xed paraforlmadehyde 4% for 15 minute at room temperature, washed 3 times with PBS, permeabilized with 0.5% Triton X-100 for ten minutes at room temperature, washed 3 times with PBS and incubated with FITC-phalloidin (1:100) for 30 minutes at RT. An LSM510 Zeiss confocal laser-scanning unit (Carl Zeiss, Germany) was used for detection.

CHAPTER IV

TABLE 2. Sequence of siRNA oligonucleotides

GENE NAME OLIGO-SENSE(5'-3') OLIGO-ANTISENSE (5'-3') TG2 GCGUCGUGACCAACUACAAdTdT UUGUAGUUGGUCACGACGCdGdG " oligo#2 CAAGACAGGUGGAGUUGAUUU pAUCAACUCCACCUGUCUUGUU " oligo#4 AAGCUGCCGUUCUUUAAUAUU pUAUUAAAGAACGGCAGCUUUU BECLIN-1 AAGAUUGAAGACACAGGAGGC GCCUCCUGUGUCUUCAAUCUU CFTR oligo#1 CGUGUGUCUGUAAACUGAUGGCUAA UUAGCCAUCAGUUUACAGACACAGC CFTR oligo#2 CCCUUCUGUUGAUUCUGCUGACAAU AUUGUCAGCAGAAUCAACAGAAGGG CFTR oligo#3 GGCAUAGGCUUAUGCCUUCUCUUUA UAAAGAGAAGGCAUAAGCCUAUGCC SUMO-1 oligo#1 GGAUAGCAGUGAGAUU CACUUCAAA UUUGAAGUGAAUCUCACUGCUA UCC SUMO-1 oligo#2 GGUGUUCCAAUGAAUU CACUCAGGU ACCUGAGUGAAUUCAUUGGAAC ACC SUMO-1 oligo#3 GGAAGAAGAUGUGAUU GAAGUUUAU AUAAACUUCAAUCACAUCUUCU UCC

TABLE 3. Sequence of primers used for Quantitative RT-PCR

FOWARD PRIMER SEQUENCE (5'- REVERSE PRIMER SEQUENCE (5'- GENE NAME 3') 3') !-actin ATTGCCGACAGGATGCAGAA ACATCTGCTGGAAGGTGGACAG ULK1 GTCCAGCTTTGACAGTCAGT TCCCAGGACTCAGGATTCCA ULK2 TGTTTCTAGGGTGTAGCTGG CACCATCTACCCCAAGACCT ATG5 AGTATCAGACACGATCATGG TGCAAAGGCCTGACACTGGT ATG6\BECLIN1 TGTCACCATCCAGGAACTCA CTGTTGGCACTTTCTGTGGA ATG7 CACTGTGAGTCGTCCAGGAC CGCTCATGTCCCAGATCTCA ATG8 CGTGGAGTCCGCGAAGATTC AGGCATGAGGACAATGCACA ATG14L ATGAGCGTCTGGCAAATCTT CCCATCGTCCTGAGAGGTAA Bcl2 GCTGCTGGAGGCTGGGGAGA CCCCCACAGGAACCCTCCTC hVsps34 AAGCAGTGCCTGTAGGAGGA TGTCGATGAGCTTTGGTGAG hVsps15 TGGGACTGGCGGTCTGCACT CGGGACGGGAAAACGCGTCA Ambra-1 AACCCTCCACTGCGAGTTGA TCTACCTGTTCCGTGGTTCTCC

CHAPTER IV

TABLE 4. Primary antibody

ANTIBODY COMPANY phospho-p42/p44 MAP kinases CELL SIGNALING phospho-p65 CELL SIGNALING -actin (13E5) CELL SIGNALING SANTA CRUZ BIOTECNOLOGY SANTA CRUZ BIOTECNOLOGY SANTA CRUZ Ubiquitin(FL76) BIOTECNOLOGY -L-glutamyl)-L-lysine isopeptide (81DIC2) COVOLAB SANTA CRUZ BIOTECNOLOGY TG2 (CUB7402) NEOMARKER SANTA CRUZ BIOTECNOLOGY SANTA CRUZ SUMO-1 BIOTECNOLOGY BECLIN-1 ABCAM SANTA CRUZ GFP(FL) BIOTECNOLOGY SANTA CRUZ hVps15\p150 (Nterm) BIOTECNOLOGY SANTA CRUZ Bcl-2(C2) BIOTECNOLOGY SANTA CRUZ -6 BIOTECNOLOGY SANTA CRUZ N-Cor BIOTECNOLOGY LC3 ABCAM CFTR ABCAM AMBRA-1 ABCAM p62 BD BIOSCIENCE BD BIOSCIENCE -TUBULIN SIGMA hVps34 SIGMA Atg14L ABCAM LAMP1 ABCAM EEA-1 ABCAM

CHAPTER IV

REFERENCES

1. Ridley A, Schwartz M, Burridge K, Firtel R, Ginsberg M, Borisy G, Parsons J, Horwitz A. Cell migration: integrating signals from front to back. Science 2003, 302(5651):1704-1709.

2. Friedl P, Gilmour D. Collective cell migration in morphogenesis, regeneration and cancer. Nat Rev Mol Cell Biol 2009, 10(7):445-457.

3. Birchmeier C, Birchmeier W, Gherardi E, Vande Woude GF. Met, metastasis, motility and more. 2003 Nat Rev Mol Cell Biol.;4(12):915-25

4. Linder S. The matrix corroded: podosomes and invadopodia in extracellular matrix degradation. Trends Cell Biol 2007, 17(3):107-117

5. Destaing O, Saltel F, Géminard J, Jurdic P, Bard F. Podosomes display actin turnover and dynamic self-organization in osteoclasts expressing actin-green fluorescent protein. Mol Biol Cell 2003, 14(2):407-416.

6. Hynes R. Integrins: bidirectional, allosteric signaling machines. Cell 2002, 110(6):673-687

7. Friedl P, Bröcker E. The biology of cell locomotion within three-dimensional extracellular matrix. Cell Mol Life Sci 2000, 57(1):41-64.

8. Geiger B, Bershadsky A, Pankov R, Yamada K. Transmembrane crosstalk between the extracellular matrix--cytoskeleton crosstalk. Nat Rev Mol Cell Biol 2001, 2(11):793-805.

9. Zamir E, Geiger B. Molecular complexity and dynamics of cell-matrix adhesions. J Cell Sci 2001, 114(Pt 20):3583-3590.

10.Legrand C, Gilles C, Zahm J, Polette M, Buisson A, Kaplan H, Birembaut P, Tournier J: Airway epithelial cell migration dynamics. MMP-9 role in cell- extracellular matrix remodeling. J Cell Biol 1999, 146(2):517-529.

11. Golubovskaya V, Cance W: Focal adhesion kinase and p53 signal transduction pathways in cancer. Front Biosci 2010, 15:901-912.

12.McEver RP. Rolling back neutrophil adhesion. Nat Immunol. 2010;11(4):282-4

CHAPTER IV

13.Entschladen F, Drell Tt, Lang K, Masur K, Palm D, Bastian P, Niggemann B, Zaenker K: Analysis methods of human cell migration. Exp Cell Res 2005, 307(2):418-426.

14.Ratjen, F., and G. Doring.. Cystic [brosis. Lancet 2003. 361: 681–689.

15. Osika, E., J. M. Cavaillon, K. Chadelat, M. Boule, C. Fitting, G. Tournier, and A. Clement. Distinct sputum cytokine pro[les in cystic [brosis and other chronic in\ammatory airway disease. Eur. Respir. J. 1999, 14: 339–346.

16. Dubin, P. J., F. McAllister, and J. K. Kolls. Is cystic [brosis a TH17 disease? In\amm. Res. 2007, 56: 221–227.

17. Escotte, S., O. Tabary, D. Dusser, C. Majer-Teboul, E. Puchelle, and J. Jacquot. Fluticasone reduces IL-6 and IL-8 production of cystic [brosis bronchial epithelial cells via IKK-f kinase pathway. Eur. Respir. J. 2003, 21: 574–581.

18.Rottner, M., C. Kunzelmann, M. Mergey, J. M. Freyssinet, and M. C. Martinez. Exaggerated apoptosis and NF-KB activation in pancreatic and tracheal cystic [brosis cells. FASEB J. 2007, 21: 2939–2948.

19.Venkatakrishnan, A., A. A. Stecenko, G. King, T. R. Blackwell, K. Brigham, J. W. Christman, and T. S. Blackwell. Exaggerated activation of nuclear factor- KB and altered IKB-beta processing in cystic [brosis bronchial epithelial cells. Am. J. Respir. Cell Mol. Biol. 2000, 23: 396 –403.

20.Becker, M. N., M. S. Sauer, M. S. Muhlebach, A. J. Hirsh, Q. Wu, M. W. Verghese, and S. H. Randell. Cytokine secretion by cystic [brosis airway epithelial cells. Am. J. Respir. Crit. Care Med. 2004, 169: 645–653.

21.Hybiske, K., Z. Fu, C. Schwarzer, J. Tseng, J. Do, N. Huang, and T. E. Machen. Effects of cystic [brosis transmembrane conductance regulator and 2+ bF508CFTR on in\ammatory response, ER stress, and Ca of airway epithelia. Am. J. Physiol 2007,. 293: L1250–L1260.

22. Perez, A., A. C. Issler, C. U. Cotton, T. J. Kelley, A. S. Verkman, and P. B. Davis. CFTR inhibition mimics the cystic [brosis in\ammatory pro[le. Am. J. Physiol. 2007, 292: L383–L395.

CHAPTER IV

23.Verhaeghe, C., C. Remouchamps, B. Hennuy, A. Vanderplasschen, A. Chariot, S. Tabruyn, C. Oury, and V. Bours.. Role of IKK and ERK pathways in intrinsic in\ammation of cystic [brosis airways. Biochem. Pharmacol. 2007, 73: 1982–1994.

24.Maiuri, L., A. Luciani, I. Giardino, V. Raia, V. R. Villella, M. D’Apolito, M. Pettoello-Mantovani, S. Guido, C. Ciacci, M. Cimmino,Cexus ON, Londei M, Quaratino S. Tissue transglutaminase activation modulates inflammation in cystic [ -regulation. J. Immunol. 2008, 180: 7697–7705.

25.Vij, N., S. Mazur, and P. L. Zeitlin. CFTR is a negative regulator of NFKB mediated innate immune response. PLoS ONE 2009, 4: e4664.

26.Sollid LM. Coeliac disease: dissecting a complex in\ammatory disorder. Nat Rev Immunol 2002;2:647e55.

27.Sollid LM. Molecular basis of coeliac disease. Annu Rev Immunol 2000;18:53e81.

28.Shan L, Molberg O, Parrot I,. Structural basis for gluten intolerance in celiac sprue. Science 2002;297:2275e79.

29.Arentz-Hansen H, Korner R, Molberg O, Quarsten H, Vader W, Kooy YM, Lundin KE, Koning F, Roepstorff P, Sollid LM, McAdam SN. The intestinal T cell response to alpha-gliadin in adult celiac disease is focused on a single deamidated glutamine targeted by tissue transglutaminase. J Exp Med 2000;191:603–612.

30.van de Wal Y, Kooy Y, Van Veelen P, Pena S, Mearin L, Papadopoulos G, Koning F. Selective deamidation by tissue transglutaminase strongly enhances gliadin-specific T cell reactivity. J Immunol 1998;161:1585–1588.

31.Kim CY, Quarsten H, Bergseng E, Khosla C, Sollid LM. Structural basis for HLA-DQ2-mediated presentation of gluten epitopes in celiac disease. Proc Natl Acad Sci U S A 2004;101:4175–4179.

32.Reinke Y, Zimmer KP, Naim HY Toxic peptides in Frazer's fraction interact with the actin cytoskeleton and affect the targeting and function of intestinal proteins. Exp Cell Res. 2009;15;315(19):3442-52.

CHAPTER IV

33.Malorni, W., M. G. Farrace, C. Rodolfo, and M. Piacentini. Type 2 transglutaminase in neurodegenerative diseases: the mitochondrial connection. Curr. Pharm. Des. 2008; 14: 278–288.

34.Beninati S, Piacentini M. The transglutaminase family: an overview: minireview article Amino Acids. 2004;26(4):367-72

35.Maiuri, L., C. Ciacci, I. Ricciardelli, L. Vacca, V. Raia, A. Rispo, M. Grif[n, T. Issekutz, S. Quaratino, and M. Londei. Unexpected role of surface transglutaminase type II in celiac disease. Gastroenterology 2005;129: 1400 – 1413.

36.Siegel, M., and C. Khosla. Transglutaminase 2 inhibitors and their therapeutic role in disease states. Pharmacol. Ther. 2007;115: 232–245.

37.Skill, N. J., T. S. Johnson, I. G. Coutts, R. E. Saint, M. Fisher, L. Huang, A. M. El Nahas, R. J. Collighan, and M. Grif[n. Inhibition of transglutaminase activity reduces extracellular matrix accumulation induced by high glucose levels in proximal tubular epithelial cells. J. Biol. Chem. 2004; 279: 47754 – 47762.

38.Yi, S. J., K. H. Kim, H. J. Choi, J. O. Yoo, H. I. Jung, J. A. Han, Y. M. Kim, I. B. (2+) Suh, and K. S. Ha. [Ca ]-dependent generation of intracellular reactive oxygen species mediates maitotoxin-induced cellular responses in human umbilical vein endothelial cells. Mol. Cells 2006; 21: 121–128.

39.Starosta, V., E. Rietschel, K. Paul, U. Baumann, and M. Griese. Oxidative changes of bronchoalveolar proteins in cystic [brosis. Chest 2006; 129: 431– 437.

40.Lorand, L., and R. M. Graham. Transglutaminases: crosslinking enzymes with pleiotropic functions. Nat. Rev. Mol. Cell Biol. 2003; 4: 140–156.

41.Meulmeester, E., and F. Melchior. Cell biology: SUMO. 2008; Nature 452: 709–711.

42.Geiss-Friedlander, R., and F. Melchior. Concepts in sumoylation: a decade on. Nat. Rev. Mol. Cell Biol. 2007; 8: 947–956

CHAPTER IV

43.Mabb, A. M., and S. M. Wuerzberger-Davis. PIASy mediates NEMO sumoylation and NF-KB activation in response to genotoxic stress. Nat. Cell Biol. 2006; 8: 986–993.

44.Mu¨ller, S., and C. Hoege. SUMO-1, ubiquitin’s mysterious cousin. Nat. Rev. Mol. Cell Biol. 2001; 2: 202–210.

45.Daynes, R. A., and D. C. Jones. Emerging roles of PPARs in in\ammation and immunity. Nat. Rev. Immunol. 2002; 2: 748–759.

46.Belvisi, M. G., D. J. Hele, and M. A. Berrel.. Peroxisome proliferatoractivated receptor y agonists as therapy for chronic airway in\ammation. Eur. J. Pharmacol. 2006; 533: 101–109.

47.Pascual, G., A. L. Fong, S. Ogawa, A. Gamliel, A. C. Li, V. Perissi, D. W. Rose, T. M. Willson, M. G. Rosenfeld, and C. K. Glass. A SUMOylation-dependent pathway mediates transrepression of in\ammatory response genes by PPAR-. Nature 2005; 437: 759–763

48.Kim, D. S., S. S. Park, B. H. Nam, I. H. Kim, and S. Y. Kim. Reversal of drug resistance in breast cancer cells by transglutaminase 2 inhibition and nuclear factor-KB inactivation. Cancer Res. 2006; 66: 10936–10943

49.Kopito, R. R. Aggresomes, inclusion bodies and protein aggregation. Trends Cell Biol. 2000; 10: 524–530.

50.Kawaguchi, Y., J. J. Kovacs, A. McLaurin, J. M. Vance, A. Ito, and T. P. Yao. The deacetylase HDAC6 regulates aggresome formation and cell viability in response to misfolded protein stress. Cell 2003; 115: 727–738.

51.Sha, Y., Pandit, L., Zeng, S. & Eissa, N. T. A critical role of CHIP in the aggresome pathway. Mol. Cell. Biol. 2009; 29, 116–128.

52.Bence, N. F., Sampat, R. M. & Kopito, R. R. Impairment of the ubiquitin- proteasome system by protein aggregation. Science 2001; 292, 1552–1555.

CHAPTER IV

53.Fu, L. & Sztul, E. ER-associated complexes (ERACs) containing aggregated cystic fibrosis transmembrane conductance regulator (CFTR) are degraded by autophagy. Eur. J. Cell Biol. 2009; 88, 215–226.

54.Luciani A, Villella VR, Vasaturo A, Giardino I, Raia V, Pettoello-Mantovani M, D'Apolito M, Guido S, Leal T, Quaratino S, Maiuri L. SUMOylation of tissue transglutaminase as link between oxidative stress and inflammation.J. Immunol. 2009; 183, 2775–2784.

55.Trudel S, Kelly M, Fritsch J, Nguyen-Khoa T, Thérond P, Couturier M, Dadlez M, Debski J, Touqui L, Vallée B, Ollero M, Edelman A, Brouillard F. Peroxiredoxin 6 fails to limit phospholipid peroxidation in lung from Cftr- knockout mice subjected to oxidative challenge. PLoS One 2009; 4, e6075.

56.Mizushima, N., Levine, B., Cuervo, A. M. & Klionsky, D. J. Autophagy fights disease through cellular self-digestion. Nature 2008; 28, 1069–1075.

57.Moreau, K., Luo, S. & Rubinsztein, D. C. Cytoprotective roles for autophagy. Curr. Opin. Cell Biol. 2010; 22, 206–211.

58.Korolchuk, V. I., Mansilla, A., Menzies, F. M. & Rubinsztein, D. C. Autophagy inhibition compromises degradation of ubiquitin–proteasome pathway substrates. Mol. Cell 2009; 33, 517–527.

59.Kirkin, V., McEwan, D. G., Novak, I. & Dikic, I. A role for ubiquitin in selective autophagy. Mol. Cell 2009; 34, 259–269.

60.Sinha, S. & Levine, B. The autophagy effector Beclin 1: a novel BH3-only protein. Oncogene 2008; 27, S137–S148.

61. Maiuri, M. C., Criollo, A. & Kroemer, G. Crosstalk between apoptosis and autophagy within the Beclin 1 interactome. EMBO J. 2010; 29, 515–516.

62.He, C. & Levine, B. The Beclin 1 interactome. Curr. Opin. Cell Biol. 2010; 22, 140–149.

CHAPTER IV

63. Klionsky, D. J., Abeliovich, H., Agostinis, P., Agrawal, D. K. & Aliev, G. Guidelines for the use and interpretation of assays for monitoring autophagy in higher eukaryotes. Autophagy 2008; 4,151–175.

64. Mizushima, N., Yoshimori, T. & Levine, B. Methods in mammalian autophagy research. Cell 2010; 140, 313–326.

65.Matsunaga, K., Saitoh T, Tabata K, Omori H, Satoh T, Kurotori N, Maejima I, Shirahama-Noda K, Ichimura T, Isobe T, Akira S, Noda T, Yoshimori l. Two Beclin 1-binding proteins, Atg14L and Rubicon, reciprocally regulate autophagy at different stages. Nature Cell Biol. 2009; 11, 385–396.

66. Zhong, Y., Wang QJ, Li X, Yan Y, Backer JM, Chait BT, Heintz N, Yue Z. Distinct regulation of autophagic activity by Atg14L and Rubicon associated with Beclin 1-phosphatidylinositol-3-kinase complex. Nature Cell Biol. 2009; 11, 468–476.

67.Axe EL, Walker SA, Manifava M, Chandra P, Roderick HL, Habermann A, Griffiths G, Ktistakis NT. Autophagosome formation from membrane compartments enriched in phosphatidylinositol 3-phosphate and dynamically connected to the endoplasmic reticulum. J. Cell Biol. 2008; 182, 685–701

68.Hayashi-Nishino, M., Fujita N, Noda T, Yamaguchi A, Yoshimori T, Yamamoto A. A subdomain of the endoplasmic reticulum forms a cradle for autophagosome formation. Nature Cell Biol. 2009; 11, 1433–1437.

69.Maiuri MC, Le Toumelin G, Criollo A, Rain JC, Gautier F, Juin P, Tasdemir E, Pierron G, Troulinaki K, Tavernarakis N, Hickman JA, Geneste O, Kroemer G. Functional and physical interaction between Bcl-XL and a BH3-like domain in Beclin-1. EMBO J. 2007; 26, 2527–2539.

70.Pattingre S, Tassa A, Qu X, Garuti R, Liang XH, Mizushima N, Packer M, Schneider MD, Levine B Bcl-2 antiapoptotic proteins inhibit Beclin 1- dependent autophagy. Cell 2005; 23, 927–939 .

CHAPTER IV

71.Liang, C., Lee JS, Inn KS, Gack MU, Li Q, Roberts EA, Vergne I, Deretic V, Feng P, Akazawa C, Jung JU. Beclin1-binding UVRAG targets the class C Vps complex to coordinate autophagosome maturation and endocytic trafficking. Nature Cell Biol. 2008; 10, 776–787.

72.Raia V, Maiuri L, Ciacci C, Ricciardelli I, Vacca L, Auricchio S, Cimmino M, Cavaliere M, Nardone M, Cesaro A, Malcolm J, Quaratino S, Londei M. Inhibition of p38 mitogen activated protein kinase controls airway in\ammation in cystic [brosis. Thorax 2005; 60: 773–780.

73.Legssyer, R., F. Huaux, J. Lebacq, M. Delos, E. Marbaix, P. Lebecque, D. Lison, B. J. Scholte, P. Wallemacq, and T. Leal. 2006. Azithromycin reduces spontaneous and induced in\ammation in bF508 cystic [brosis mice. Respir. Res. 7: 134

74.Spencer, B., Potkar R, Trejo M, Rockenstein E, Patrick C, Gindi R, Adame A, Wyss-Coray T, Masliah E. Beclin 1 gene transfer activates autophagy and -synuclein models of Parkinson’s and Lewy body diseases. J. Neurosci. 2009; 29, 13578–13588.

75.Stocker, A. G., Kremer KL, Koldej R, Miller DS, Anson DS, Parsons DW. Single-dose lentiviral gene transfer for lifetime airway gene expression. J. Gene Med. 11, 861–867 (2009).

76.Stark HJ, Szabowski A, Fusenig NE, Maas-Szabowski N. Organotypic cocultures as skin equivalents: A complex and sophisticated in vitro system. 2004 Biol Proced Online.;6:55-60.

77. Maiuri L, Ciacci C, Ricciardelli I, Vacca L, Raia V, Auricchio S, Picard J, Osman M, Quaratino S, Londei M. Association between innate response to gliadin and activation of pathogenic T cells in coeliac disease. Lancet 2003;362:30e7

78.Barone MV, Gimigliano A, Castoria G, Paolella G, Maurano F, Paparo F, Maglio M, Mineo A, Miele E, Nanayakkara M, Troncone R, Auricchio S. Growth factor-like activity of gliadin, an alimentary protein: implications for coeliac disease. Gut 2007;56:480e8

CHAPTER IV

79.Matysiak-Budnik T, Candalh C, Dugave C, et al. Alterations of the intestinal transport and processing of gliadin peptides in celiac disease. Gastroenterology 2003;125:696e707.

80.Lu A, Tebar F, Alvarez-Moya B, et al. A clathrin-dependent pathway leads to KRas signaling on late endosomes en route to lysosomes. J Cell Biol 2009;6:863e79.

81.Taylor, J. P., F. Tanaka, J. Robitschek, C. M. Sandoval, A. Taye, S. Markovic- Plese, and K. H. Fischbeck. Aggresomes protect cells by enhancing the degradation of toxic polyglutamine containing protein. Hum. Mol. Genet. 2003; 12: 749- 757.

82.Johnston, J. A., C. L. Ward, and R. R. Kopito. Aggresomes: a cellular response to misfolded proteins. J. Cell Biol. 1998;143: 1883–1898.

83.Maiuri, C., Zalckvar, E., Kimchi, A. & Kroemer, G. Self-eating and self-killing: crosstalk between autophagy and apoptosis. Nature Rev Mol. Cell. Biol. 2007; 8, 741–752.

84.Hara, T., Nakamura K, Matsui M, Yamamoto A, Nakahara Y, Suzuki- Migishima R, Yokoyama M, Mishima K, Saito I, Okano H, Mizushima N. Suppression of basal autophagy in neural cells causes neurodegenerative disease in mice. Nature 2006; 441, 885–889.

85.Tebbenkamp, A. T. & Borchel, D. R. Protein aggregate characterization in models of neurodegenerative disease. Methods Mol. Biol. 2009; 566, 85–91.

86.Martínez, A, Portero-Otin, M., Pamplona, R. & Ferrer, I. Protein targets of oxidative damage in human neurodegenerative diseases with abnormal protein aggregates. Brain Pathol. 2010; 20, 281–297.

87.Luciani A, Villella VR, Esposito S, Brunetti-Pierri N, Medina D, Settembre C, Gavina M, Pulze L, Giardino I, Pettoello-Mantovani M, D'Apolito M, Guido S, Masliah E, Spencer B, Quaratino S, Raia V, Ballabio A, Maiuri L. Defective CFTR induces aggresome formation and lung inflammation in cystic fibrosis through ROS-mediated autophagy inhibition. Nat Cell Biol. 2010;12(9):863-75

CHAPTER IV

88.Luciani A, Villella VR, Vasaturo A, Giardino I, Pettoello-Mantovani M, Guido S, Cexus ON, Peake N, Londei M, Quaratino S, Maiuri L. Lysosomal accumulation of gliadin p31-43 peptide induces oxidative stress and tissue transglutaminase-mediated PPARgamma downregulation in intestinal epithelial cells and coeliac mucosa. Gut. 2010;59(3):311-9.

CHAPTER IV

APPENDIX

Publications

Luciani A & Villella VR, Esposito S, Brunetti-Pierri N, Medina D, Settembre C, Gavina M, Pulze L, Giardino I, Pettoello-Mantovani M, D'Apolito M, Guido S, Masliah E, Spencer B, Quaratino S, Raia V, Ballabio A, Maiuri L. Defective CFTR induces aggresome formation and lung inflammation in cystic fibrosis through ROS-mediated autophagy inhibition. Nat Cell Biol. 2010;12(9):863-75

Maiuri, L., A. Luciani, I. Giardino, V. Raia, V. R. Villella, M. D’Apolito, M. Pettoello-Mantovani, S. Guido, C. Ciacci, M. Cimmino,Cexus ON, Londei M, Quaratino S. Tissue transglutaminase activation modulates \in cystic [ -regulation. J. Immunol. 2008, 180: 7697–7705.

Experiences in foreign laboratories

University of Southampton School of Medicine, Cancer Science Division, General Hospital, Southampton, UK. Scientific collaboration for my PhD project, under supervision of Prof.ssa Sonia Quaratino and Dott. Nick Peake and in collaboration with Prof. Luigi Maiuri of European Institute for Research in Cystic Fibrosis and Prof. S. Guido of the University of Naples (Oct 2009-Jan 2010).

ARTICLES

Defective CFTR induces aggresome formation and lung inflammation in cystic fibrosis through ROS-mediated autophagy inhibition

Alessandro Luciani1,2,11, Valeria Rachela Villella2,11, Speranza Esposito2,3, Nicola Brunetti-Pierri4,5, Diego Medina4, Carmine Settembre4, Manuela Gavina1, Laura Pulze1, Ida Giardino6, Massimo Pettoello-Mantovani3, Maria D’Apolito3, Stefano Guido2, Eliezer Masliah7, Brian Spencer7, Sonia Quaratino8, Valeria Raia9,12, Andrea Ballabio4,10,12 and Luigi Maiuri1,3,12

Accumulation of unwanted/misfolded proteins in aggregates has aggregates after overload of polyubiquitylated proteins3–5. The accumula- been observed in airways of patients with cystic fibrosis (CF), tion of unwanted/misfolded protein aggregates has been described in sev- a life-threatening genetic disorder caused by mutations in the eral human disorders, such as neurodegenerative diseases6,7, myopathies8 gene encoding the cystic fibrosis transmembrane conductance and cancer9. Recently we described the presence of peroxisome proliferator- regulator (CFTR). Here we show how the defective CFTR activated receptor (PPAR)-γ in aggresomes in CF airways10. CF is caused results in defective autophagy and decreases the clearance of by mutations in the CFTR gene11 and the mutant CFTRF508del has been aggresomes. Defective CFTR-induced upregulation of reactive described as the prototype of a misfolded aggresome-prone protein12–14. oxygen species (ROS) and tissue transglutaminase (TG2) Sequestration of the anti-inflammatory PPAR-γ into aggresomes, drive the crosslinking of beclin 1, leading to sequestration mediated by the increased TG2 levels, contributes to the lung inflam- of phosphatidylinositol-3-kinase (PI(3)K) complex III and mation in CF15. This is the result of ROS-mediated TG2 SUMOylation accumulation of p62, which regulates aggresome formation. inducing the crosslinking, ubiquitylation and aggresome sequestration Both CFTR knockdown and the overexpression of green of several target proteins15. Because oxidative stress and accumulation fluorescent protein (GFP)-tagged-CFTRF508del induce beclin 1 of protein aggregates are features of CF airway epithelia10,15,16, we spec- downregulation and defective autophagy in non-CF airway ulated that the accumulation of aggresomes in CF is due to defective epithelia through the ROS–TG2 pathway. Restoration of beclin 1 autophagy. and autophagy by either beclin 1 overexpression, cystamine or Here we demonstrate how defective CFTR drives autophagy inhibi- antioxidants rescues the localization of the beclin 1 interactome tion in mouse and human CF airway epithelia by means of the ROS–TG2 to the endoplasmic reticulum and reverts the CF airway pathway, with sequestration of beclin 1 interactome, a key player of phenotype in vitro, in vivo in Scnn1b-transgenic and CftrF508del autophagosome formation17–19, in the aggresomes. Rescuing beclin 1 homozygous mice, and in human CF nasal biopsies. Restoring and autophagy restores CFTR trafficking and damps down inflamma- beclin 1 or knocking down p62 rescued the trafficking of tion. Our results provide evidence that defective autophagy has a critical CFTRF508del to the cell surface. These data link the CFTR defect function in CF pathogenesis and suggest that drugs restoring beclin 1 to autophagy deficiency, leading to the accumulation of protein and autophagy are potentially effective in the treatment of CF. aggregates and to lung inflammation. RESULTS Autophagy is a cytoprotective mechanism for the degradation of misfolded/ Human and mouse CF airway epithelia have defective autophagy polyubiquitylated proteins and damaged organelles through lysosome- We cultured human CF airway epithelial IB3-1 or CFBE41o– cells, car- mediated self-digestion1,2. Autophagy is important in clearing protein rying F508del/W1282X and F508del/F508del mutations, respectively,

1European Institute for Research in Cystic Fibrosis, San Raffaele Scientific Institute, Milan 20132, Italy. 2Department of Chemical Engineering, Federico II University, Naples 80125, Italy. 3Institute of Pediatrics, University of Foggia, Foggia 71100, Italy. 4Telethon Institute of Genetics and Medicine (TIGEM), Naples 80131, Italy. 5Department of Pediatrics, Federico II University, Naples 80131, Italy. 6Department of Laboratory Medicine, University of Foggia, Foggia 71100, Italy. 7Departments of Neurosciences, University of California, San Diego, La Jolla, CA 92093-0624, USA. 8Cancer Research UK Oncology Unit, University of Southampton, Southampton, SO16 6YD, U.K. 9Cystic Fibrosis Unit, Department of Pediatrics, Federico II University, Naples 80131, Italy. 10Medical Genetics, Department of Pediatrics, Federico II University, Naples 80131, Italy. 11These authors contributed equally to this work 12Correspondence should be addressed to L.M., A.B. or V.R. (email: [email protected]; [email protected]; [email protected])

Received 6 May 2010; accepted 29 July 2010; published online 15 August 2010; DOI:10.1038/ncb2090.

ab Starvation – + 12 C38 IB3-1 * N 10

C38 N

C38 8

6 * GFP–LC3

N N 4 IB3-1 2

100 C38 IB3-1 IB3-1 Number of autophagosomes per cell 0 Starvation – + 80 * N

60 f C38 IB3-1 d C38 IB3-1 Mr(K) * M (K) 40 LC3 I r 250 LC3 II 17 130 dots (percentage) 20 90 Cells with >5 GFP–LC3 Actin 36 72 0 36 Starvation – +

e 17 IB: ubiquitin c E64d/ C38 pepstatinA –+M (K) 1,000 r 3.3 % LC3 I 11 500 LC3 II 117 Count Actin 0 Actin 1,000 36 IB3-1 IB3-1 15.4 % 500 i

Count Human nasal mucosa 10 Human nasal mucosa 0 h Control CF M (K) r 8 Control of tissue p62 2 CF g C38 IB3-1 M (K) r 55 6 p62 LC3 I 55 LC3 II 17 4 Actin 55 36 Tubulin 2 *

k LC3+ dots per mm 0 10

Mice Mice Human nasal mucosa 8 F508del j Control CF l WT CFTR 6 WT L CFTRF508del 4 p62 p62 * 2 L

LC3+ dots per mm2 of tissue 0

Figure 1 Autophagy is defective in human and mice CF airway epithelia. C38 Information, Fig. S10. (h) Immunoblot of LC3 and p62 in control (n = 10) and IB3-1 cells transfected with GFP–LC3 plasmid in nutrient-rich or starvation and CF (n = 10) human nasal mucosa. (i) Number of LC3-positive dots medium. (a) Top: confocal microscopy of GFP. N, nucleus. Scale bar, 10 μm. (per 104 mm–2 of mucosa) in control (n = 10) and CF (n = 10) human nasal Bottom: percentage of cells containing more than five GFP–LC3 punctate dots mucosa. Means ± s.d. Asterisk, P < 0.01 versus controls; ANOVA. (j) Confocal per cell. Means ± s.d. (n = 5). Asterisk, P < 0.001 versus C38 cells; analysis microscopy of p62 in control and CF human nasal mucosa. Representative of variance (ANOVA). (b) Left: electron microscopy of starved C38 and IB3-1 images from ten patients with CF. Nuclei counterstained with 4ʹ,6-diamidino- cells. The right-hand images are higher magnifications showing areas enriched 2-phenylindole (DAPI; blue). Scale bar, 10 μm. (k) Number of LC3-positive in autophagosomes in C38 cells (arrows) and perinuclear mitochondria in IB3-1 dots in lung tissues from wild-type (WT) and CftrF508del mice. Means ± s.d. for cells. N, nucleus. Scale bar, 250 nm. Right: numbers of autophagosomes per lung tissues from ten mice per group. Asterisk, P < 0.001 versus WT mice; cell counted in 20 cells per experiment. Means ± s.d. for three independent ANOVA. (l) Confocal microscopy of p62 in WT and CftrF508del homozygous mice. experiments. Asterisk, P < 0.001 versus C38 cells; ANOVA. (c) Fluorescence- Representative staining from at least five airway epithelial areas per mouse activated cell sorting (FACS) analysis after staining with fluorescent dye DiOC6 (n = 10 mice per group). L, lumen. Nuclei counterstained with DAPI (blue). in starved C38 and IB3-1 cells. (d, e) Immunoblot of LC3 in starved C38 and Scale bar, 10 μm. β-Actin was used as loading control for immunoblot analysis IB3-1 cells (d) and IB3-1 cells (e) after incubation with E64d and pepstatin of cells, and αβ-tubulin for nasal mucosa. All expreriments were repeated three A. (f, g) Immunoblots (IB) of polyubiquitin (f) and p62 (g) in starved C38 times. Densitometric analysis of blots is shown in Supplementary Information, and IB3-1 cells. Uncropped images of blots are shown in Supplementary Fig. S9.

a + starvation b + starvation c + starvation d + starvation C38 IB3-1 C38 IB3-1 IB3-1 C38 IB3-1 Mr(K) Ambra 1 130 N N N N N N hVps15 hVps34

HDAC6 HDAC6 130 hVps34

hVps34 95 Atg14L 55 N N N N N Beclin 1 N hVps34 Beclin 1 55 Beclin 1 Calnexin Actin 36

N N

Merge N N N Merge

Merge N Merge

ef+ HA-beclin 1 g LC3 Empty vector Empty vector HA–beclin 1 Empty vector HA Calnexin Merge HA–beclin 1 HA–beclin 1

100 20 NNN N 80 * * 15

Empty vector 60 hVps34 Calnexin Merge 10 40 per cell 5 (percentage) (percentage) IB3-1 N 20 NNN 0 0 Cells with >5 LC3 dots

HA–beclin 1 IB3-1 IB3-1 IB3-1 Number of autophagosomes IB3-1; + starvation

hiEmpty HA– Empty HA– j kl+LV-GFP F508del +LV-GFP +LV-beclin 1 vector beclin 1 vector beclin 1 WT Cftr M (K) 2 +LV-beclin 1 M (K) M (K) r L HA r r Beclin-1 10 p62 55 * L 55 8

LC3 I p62 Actin 55 6 LC3 II 17 36 Tubulin 4

IB3-1 of tissue 2 CftrF508del mice Actin mice 0

36 LC3+ dots per mm CftrF508del mice IB3-1

Figure 2 Sequestration of beclin 1 interactome in aggresomes drives defective IB3-1 cells shows areas enriched in autophagosomes (arrows). Scale bar, autophagy in CF airway epithelia. (a–d) Starved C38 and IB3-1 cells. (a–c) 250 nm. Right: number of autophagosomes per cell; 20 cells were counted Confocal microscopy images of beclin 1 (green) and hVps34 (red) (a), calnexin in each experiment. Means ± s.d. for three independent experiments. (green) and hVps34 (red) (b), and beclin 1 (green) and HDAC6 (red) (c, left) Asterisk, P < 0.001 versus cells transfected with the empty vector; ANOVA. or hVps34 (green) and HDAC6 (red) (c, right). N, nucleus. Scale bar, 10 μm. (h) Immunoblot of HA and LC3. (i) Immunoblot of p62. Uncropped images Yellow indicates co-localization. (d) Immunoblot of Ambra1, hVps15, hVps34, of blots are shown in Supplementary Information, Fig. S10. (j) Immunoblot Atg14L and beclin 1. Uncropped images of blots are shown in Supplementary of beclin 1 in WT and CftrF508del mice. Representative of seven mice for each Information, Fig. S10. (e–i) Starved IB3-1 cells transfected with either HA- group. (k, l) CftrF508del mice (n = 3 per group) were administered intranasally tagged human beclin 1 or the empty vector. (e) Confocal microscopy images with a lentiviral vector encoding beclin 1 (LV-beclin 1) or GFP (LV-GFP). (k) of HA (red) and calnexin (green) (top) or hVps34 (red) and calnexin (green) LC3-positive dots per mm2 of lung tissues. Means ± s.d. for lung tissues from (bottom). Yellow indicates co-localization. N, nucleus. Scale bar, 10 μm. (f) three mice per group. (l) Confocal microscopy micrographs of p62. L, lumen. Left: confocal images of LC3-positive dots. N, nucleus. Scale bar, 10 μm. Representative stainings from at least five airway epithelial areas per mouse. Right: percentage of cells containing more than five LC3-positive dots. At least Scale bar, 10 μm. β-Actin was used as loading control for immunoblot analysis 60 cells were counted in each experiment. Means ± s.d. for three independent of cells, and αβ-tubulin for mouse tissue. Densitometric analysis of blots and experiments. Asterisk, P < 0.001 versus cells transfected with the empty quantitative measurement of co-localizations are shown in Supplementary vector; ANOVA. (g) Left: electron microscopy of HA–beclin 1-transfected Information, Fig. S9. and the isogenic stably rescued C38, S9 (refs 10, 15) or bronchial epi- E64d and pepstatin A, inhibitors of the lysosomal proteases (Fig. 1e), thelial 16HBE14o– cells10,15, under nutrient starvation to stimulate and also chloroquine, a weak base disrupting lysosomal functions21 autophagy20,21. Transfection of GFP–LC3, a marker of autophago- (Supplementary Information, Fig. S1b), failed to increase LC3 II, thus somes21, revealed a larger number of GFP–LC3 puncta in C38 cells indicating that decreased LC3 II levels were not due to lysosomal deg- than in IB3-1 cells (P < 0.001) (Fig. 1a). Transmission electron micro- radation of LC3. scopy showed autophagosomes containing mitochondria and cellular Defective autophagy in IB3-1 cells was also revealed by the increased organelles in C38 cells but rarely in IB3-1 cells (Fig. 1b; Supplementary pool of polyubiquitylated proteins (Fig. 1f) and the accumulation of Information, Fig. S1a). Increased mitochondrial membrane depolari- p62 (Fig. 1g; Supplementary Information, Fig. S1c), a LC3-binding and zation was observed in IB3-1 cells but not in C38 cells (Fig. 1c). ubiquitin-binding protein, that accumulates in intracellular aggregates An increase in LC3 II, the lipidated product of LC3 (refs 20, 21), was under defective autophagy3,20,21. IB3-1 cells were also resistant to rapamy- observed in C38 cells but not in IB3-1 cells under starvation (Fig. 1d). cin, a known inducer of autophagy20,21 (Supplementary Information,

c ab HDAC6 Beclin 1 Merge MG132 –+ M (K) M (K) r r NN N 250 Control sirRNATG2 siRNACystamine 250

130 IB: isopeptide 130 NN N IB: isopeptide Beclin 1 Beclin 1 IP: beclin 1 IB3-1 + TG2 siRNA siRNA + control IP: beclin 1 IB3-1 IB3-1; +MG132 g d e Control TG2 f Control siRNA+MG132 Control siRNA IB3-1 siRNA siRNA M (K) TG2 siRNA+MG132 r PIASy siRNA TG2 72 Cystamine Al546, beclin 1 100 EUK-134 Ambra1 Cy5, HDAC6 Empty vector WT-TG2 C277S- TG2 140 130 Mr(K) 55 80 * 130 hVps15 130 Beclin 1 * C38 * 120 hVps34 Actin 60 95 36 110 Atg14L Beclin 1 55 40 ) Al546 (a.u.)

0 55 IB3-1 100 Beclin 1 55 Actin ( F/F 36 dots (percentage) 20

90 Actin Cells with >5 GFP-LC3 0 204060 36 0 Time frame (s) IB3-1 IB3-1

h Control siRNA PIASy siRNA EUK134 Cystamine * 10 * * 8 IB3-1 6 Control siRNA

per cell 4 PIASy siRNA Number of 2 Cystamine autophagosomes 0 EUK-134 IB3-1

ijCystamine +– M (K) r Control siRNAPIASy siRNA Cystamine EUK-134M (K) Control siRNA PIASy siRNA r Mr(K) p62 Beclin-1 55 PIASy 55 55 Actin Actin Actin 36 36 36 IB3-1 IB3-1 IB3-1

Figure 3 TG2-mediated crosslinking of beclin 1 induces aggresome of blots are shown in Supplementary Information, Fig. S10. (g, h) IB3-1 cells sequestration of beclin 1 interactome and drives defective autophagy in CF were treated with cystamine or 10 mM EUK-134 or transfected with either airway epithelial cells. (a) IB3-1 cells cultured with 250 μM cystamine or 50 nM human PIASy siRNA or scrambled oligonucleotides. (g) Percentage of transfected with either 50 nM human TG2 siRNA or scrambled oligonucleotides. cells containing more than five GFP–LC3 punctate dots. Means ± s.d. for three IP, immunoprecipitation; IB, immunoblot. (b) IB3-1 cells cultured with the independent experiments. Asterisk, P < 0.001 versus control; ANOVA. (h) Left: proteasome inhibitor MG132 (50 μM). (c, d) IB3-1 cells transfected with either electron micrographs of IB3-1 cells showing areas enriched in autophagosomes human TG2 siRNA or scrambled oligonucleotides followed by incubation with (arrows) on PIASy gene silencing, EUK-134 or cystamine. Scale bar, 100 nm. MG132. (c) Cells immunostained with anti-beclin 1 (green) and anti-HDAC-6 Right: number of autophagosomes per cell counted in 20 cells per experiment. (red) antibodies. Yellow indicates co-localization. N, nucleus. Scale bar, 10 μm. Means ± s.d. for three independent experiments. Asterisk, P < 0.001 versus (d) FRET analysis of beclin 1-Alexa 546 (Al546) fluorescence after HDAC6-Cy5 control; ANOVA. (i) IB3-1 cells treated with cystamine. Immunoblot of p62. photobleaching. F/F0 indicates the post-bleaching/pre-bleaching fluorescence (j) IB3-1 cells treated with cystamine or EUK-134 or transfected with human ratio of Al546/Cy5. (e) IB3-1 cells were transfected with either human TG2 PIASy siRNA or scrambled oligonucleotides. Immunoblot of beclin 1 (left) siRNA or scrambled oligonucleotides. Immunoblot of TG2, Ambra-1, hVps15, and PIASy (right). Uncropped images of blots are shown in Supplementary hVps34, Atg14L and beclin 1 in IB3-1 cells grown in starvation. Uncropped Information, Fig. S10. The experiments were repeated three times. β-actin images of blots are shown in Supplementary Information, Fig. S10. (f) IB3-1 was used as loading control for all blots. Densitometric analysis of blots and and C38 cells were transfected with either 1 μg of pLPCX-WT-TG or pLPCX- quantitative measurement of co-localizations are shown in Supplementary Cys277Ser-TG or the empty vector. Immunoblot of beclin 1. Uncropped images Information, Fig. S9.

Fig. S1d). CFBE41o– cells behaved in a similar manner to IB3-1 cells, of LC3 II protein and LC3 dots (per 104 mm–2 of mucosa, P < 0.01) whereas both S9 and 16HBE14o– cells behaved in a similar manner to (Fig. 1h, i), an increased pool of polyubiquitylated proteins (data not the C38 cell line (data not shown). shown), and increased levels of p62 with accumulation in intraepithe- Defective basal autophagy was observed in human and mouse CF lial aggregates in comparison with non-CF controls (n = 10) (Fig. 1h, j). airways. Nasal polyp biopsies from patients with severe CF (n = 10) A decrease in LC3 dots (per mm2 of tissue, P < 0.001), an increased (Supplementary Information, Table S1)10,15,22, showed decreased levels pool of polyubiquitylated proteins (data not shown) and intraepithelial

c abCFTR siRNA 100 C38 Control siRNA 80 ° CFTR siRNA GFP–CFTRF508del – + ° CFTR Control ° inh-172 Empty vector + – M (K) 60 r siRNA Cystamine ° Mr(K) TG2 72 p62 55 40 * PIASy Beclin 1 * 55 55 20 dots (percentage) dots (percentage) Actin LC3 I Cells with >5 GFP–LC3 0 36 LC3 II 17 A549 Actin 36 C38; + starvation +EUK-134 +cystamine d e f A549 F508del Empty vector F508del GFP–CFTR 140 GFP–CFTR F508del F508del GFP–CFTR GFP–CFTR +EUK-134 GFP–CFTRF508del +EUK-134 Empty vector 130 100 Empty

vector Cystamine 120 Mr(K) Al546, TG2 80 ** Cy5, SUMO-1 Beclin 1 55 110 * ) Al546 (a.u.) 0 LC3 I 60

( F/F 100 LC3 II 17 40 90 Actin 36 0 4020 60 A549; + starvation Time frame (s) dots (percentage) 20 Cells with >5 GFP–LC3 g p62 GFP Merge 0 A549; + starvation

NNNN N + Empty vector

F508del

NN N GFP–CFTR

Figure 4 Defective CFTR drives inhibition of autophagy by means of ROS- 250 μM cystamine. (c) Immunoblot of TG2 and PIASy. (d) FRET analysis mediated TG2 SUMOylation in airway epithelial cells. (a, b) C38 cells of TG2-Alexa 546 fluorescence after SUMO-1 Cy5 photobleaching. (e) were transfected with either 50 nM human CFTR siRNA or scrambled Immunoblot of beclin 1 and LC3 in A549 cells after starvation. Uncropped oligonucleotides (a, b) or cultured with CFTRinh-172 (b) in the presence or images of blots are shown in Supplementary Information, Fig. S10. (f) absence of cystamine (250 μM) or EUK-134 (10 mM). (a) Immunoblot of Percentage of cells containing more than five GFP–LC3 punctate dots per beclin 1, LC3 and p62 in C38 cells on starvation. Uncropped images of blots cell. Means ± s.d. for three independent experiments. Asterisk, P < 0.001; are shown in Supplementary Information, Fig. S10. (b) Percentage of cells two asterisks, P < 0.01; ANOVA. (g) Immunostaining with anti-p62 (red) containing more than five GFP–LC3 punctate dots per cell. Means ± s.d. and anti-GFP (green) antibodies in starved A549 cells. Yellow indicates for three independent experiments. Asterisk, P < 0.001 versus scrambled co-localization. N, nucleus. Scale bar, 10 μm. Each experiment was oligonucleotides; circle, P < 0.01 versus CFTR siRNA or CFTRinh-172; repeated three times. β-Actin was used as loading control for immunoblot. ANOVA. (c–g) A549 cells were transfected with 1 μg of pGFP–-CFTRF508del Densitometric analysis of blots and quantitative measurement of co- or pGFP (empty vector), in the presence or absence of 10 mM EUK-134 or localizations are shown in Supplementary Information, Fig. S9. accumulation of p62 aggregates were observed in lung tissues from was observed in CF epithelia (Supplementary Information, Table S2; CFTRF508del homozygous (CftrF508del) mice23 (n = 10) compared with their Supplementary Information, Fig. S1e, f). control littermates (n = 10) (Fig. 1k, l). A growing body of evidence indicates that beclin 1, a tumour suppressor gene17–19, interacts with the class III PI(3)K hVps34 (refs 24, 25), and Sequestration of beclin 1 interactome in aggresomes drives endoplasmic-reticulum-associated class III PI(3)K activity is crucial in defective autophagy in CF airway epithelia the initiation of autophagosome formation26,27. Moreover, dissociation We analysed the expression levels of several autophagy-related genes21,24, of beclin 1 from Bcl-2 promotes autophagy during stress conditions including ULK1, ULK2, ATG5, ATG6 (encoding beclin 1), ATG7, ATG14, such as starvation18,26–29. Bcl-2, hVps34 (encoding vacuolar protein sorting 34), hVps15 (encod- We immunoprecipitated beclin 1 from IB3-1 and C38 cells under ing regulatory myristoylated kinase) and Ambra1, a beclin 1-interact- nutrient-rich conditions or starvation. Decreased Bcl-2 immunoreactiv- ing WD40-domain protein19. Real-time PCR did not reveal decreased ity was observed in IB3-1 cells under basal conditions and was negligible messenger RNA expression levels of any of the tested genes in IB3-1 and under starvation (Supplementary Information, Fig. S2a). Fluorescence CFBE41o– cells in comparison with C38 and 16HBE14o– cells. In con- resonance energy transfer (FRET) analysis confirmed decreased interac- trast, significant increased expression of ATG7, ATG14, Bcl-2 and hVps34 tion between beclin 1 and Bcl-2 (Supplementary Information, Fig. S2b),

a + Rosiglitazone b IB3-1 IB3-1; + starvation Empty vector + HA–Beclin 1 Empty vector + rosiglitazone c Empty vector HA–Beclin 1 HA–Beclin1+ rosiglitazone Control siRNA + rosiglitazone p62 siRNA + rosiglitazone 140 N N LC3 N

PPAR- N

130

120 Al546, PPAR Cy5, N-CoR N N N CFTR N HDAC6 110

(F/F0) Al.546 (a.u.) 100

90 N N N 0204060 Merge Merge N Time frame (s) IB3-1 e ++– d LAMP-1 CFTR Merge Control siRNA Cystamine – ++ Beclin 1 siRNA – – +

N N N N N N CFTR

N N N N N N p62

N N N +3 MA Cystamine Medium N

Cystamine N N Merge Merge IB3-1; + starvation IB3-1; + starvation

f GM-130 GFP Merge g IB3-1 GFP–CFTRF508del 140 GFP–CFTRF508del + cystamine

130 N N N F508del 120 GFP Cy3, HDAC6 ) GFP (a.u.)

0 110 GFP–CFTR N NN ( F/F 100 + Cystamine 90 IB3-1 0204060 Time frame (s)

Figure 5 Restoring beclin 1 and autophagy decreases aggresome and anti-CFTR (green). Yellow indicates co-localization. N, nucleus. Scale accumulation in CF epithelia. (a, b) IB3-1 cells were transfected with bar, 10 μm. (e) IB3-1 cells were transfected with either 50 nM human either HA-tagged human beclin 1, empty vector, p62 siRNA or scrambled beclin 1 siRNA or scrambled oligonucleotides in the presence or absence oligonucleotides followed by rosiglitazone (10 μM). (a) Cells were of 250 μM cystamine under nutrient starvation. Cells were immunostained immunostained with anti-HDAC6 (red) and anti-PPAR-γ (green) antibodies. with anti-p62 (red) and anti-CFTR (green) antibodies. Yellow indicates Yellow indicates co-localization. N, nucleus. Scale bar, 10 μm. (b) FRET co-localization. N, cell nucleus. Scale bar, 10 μm. (f, g) IB3-1 cells were analysis of increase in PPAR-γ–Alexa 546 fluorescence after N-CoR Cy5 transfected with 1 μg of pGFP–CFTRF508del or the empty vector, in the photobleaching. (c) IB3-1 cells were transfected with either HA-tagged presence or absence of 250 μM cystamine. (f) Cells were immunostained human beclin 1 or the empty vector under nutrient starvation. Cells were with anti-GM-130 (red) and anti-GFP (green) antibodies. Yellow indicates co- immunostained with anti-LC-3 (red) and anti-CFTR (green). Yellow indicates localization (arrows). Scale bar, 10 μm. (g) FRET analysis of GFP fluorescence co-localization. N, nucleus. Scale bar, 10 μm. (d) IB3-1 cells were cultured after HDAC6 Cy3 photobleaching. Each experiment was repeated three times. under nutrient starvation with cystamine (250 μM) in the presence or absence Quantitative measurement of co-localizations is shown in Supplementary of the PI(3)K inhibitor 3-MA. Immunostaining was with anti-LAMP-1 (red) Information, Fig. S9. suggesting an intracellular environment favourable to the induction of on starvation (Supplementary Information, Fig. S2a). Because the PI(3) autophagy. K core complex binds to either Atg14L or the ultraviolet-irradiation- Immunoreactivities to hVps34, hVps15 and Ambra1 were detected in resistance-associated gene (UVRAG)24,30, involved in distinct functional beclin 1 immunoprecipitates from both C38 and IB3-1 cells and increased complexes driving either autophagy or endosome-to-Golgi retrograde

GFP–CFTRF508del F508del F508del a GFP–CFTRF508del b GFP–CFTR GFP–CFTR + cystamine 3.0 Cystamine ––+ – 0 μm 1 μm 2 μm 0 μm 1 μm 2 μm 2.5 Control siRNA –+– – p62 siRNA ––– + 2.0 Band B Band C Empty vector –+ – – 1.5 Mr(K) 1.0 3 μm 4 μm 5 μm 3 μm 4 μm 5 μm C 250 B 0.5

Actin Relative bands B and C 0 36 Cystamine ––+ IB3-1: cell lysate Control siRNA –– + 6 μm 7 μm 8 μm 6 μm 7 μm 8 μm p62 siRNA – – + c + cystamine

x–z

F508del d GFP–CFTR e f –– + Empty vector HA–beclin 1M (K) Control siRNA p62 siRNA Cystamine r M (K) +–– M (K) r Empty vector r TG2 72 p62 C 250 55 E-cadherin PIASy 130 55 p42/44 Actin 36 36 PPAR- 55 IP: streptavidin IB: anti-GFP Actin p42/44 36 36 IB3-1 Actin 36 IB3-1

g Cystamine –+–+ h Beclin 1 siRNA –+–+ * ) 140 * Control siRNA + –+– –1 IB3-1 M (K) 120 r 100 Control siRNA Beclin 1 siRNA p42/44 (pg ml 80

36 60 Cystamine 40 Beclin 1 siRNA Actin 20 TNF- ATG5 siRNA 36 0 IB3-1 +cystamine

Figure 6 Restoring beclin 1 and autophagy rescues CF phenotype in IB3-1 precipitated with streptavidin beads. Immunoblot of GFP to reveal the C cells. (a–d) IB3-1 cells transfected for 24 h at 37 °C with 1 μg of pGFP– form. E-cadherin and β-actin were used as positive and negative controls, CFTRF508del or empty vector in the presence or absence of 250 μM cystamine respectively. (e) IB3-1 cells transfected with either HA-tagged human or with either 50 nM human p62 siRNA or scrambled oligonucleotides. beclin 1 or empty vector. Immunoblot of PIASy, TG2, phospho-p42/44 (a) Left: CFTR revealed with anti-GFP primary antibody in IB3-1 cell and PPAR-γ proteins. (f) IB3-1 cells transfected with either 50 nM human p62 siRNA or scrambled oligonucleotides. Immunoblot of p62 and lysate. Top arrow, complex-glycosylated form (apparent Mr of band C about 170K, revealed as 210K with GFP tag); bottom arrow, core-glycosylated phospho-p42/44. Uncropped images of blots are shown in Supplementary Information, Fig. S10. (g) IB3-1 cells transfected with either 50 nM form (apparent Mr of band B about 150K, revealed as 190K with GFP tag). Uncropped images of blots are shown in Supplementary Information, human beclin 1 siRNA or scrambled oligonucleotides in the presence or Fig. S10. Right: quantification of bands B (black) and C (white) absence of cystamine. Immunoblot of phospho-p42/44. (h) IB3-1 cells normalized to β-actin levels. Values are means ± s.d. for three independent transfected with either 50 nM human beclin 1 siRNA or 50 nM human experiments. (b) Cells immunostained with an anti-GFP antibody under ATG5 siRNA or scrambled oligonucleotides in the presence or absence of non-permeabilizing conditions. Serial confocal sections were collected cystamine. TNF-α secretion. Each bar represents the mean ± s.d. for n = 3 from top to bottom of cell. (c) Side view (X–Z) of a z-stack of confocal experiments. Asterisk, P < 0.01; ANOVA. Each experiment was repeated images taken in b. (d) Surface biotinylation assays was performed with three times. β-Actin was used as loading control. Densitometric analysis membrane-impermeable sulpho-NHS-LC-biotin. The cells were fractioned of blots and quantitative measurement of co-localizations are shown in to obtain the plasma membrane fraction, and biotinylated proteins were Supplementary Information, Fig. S9. trafficking30, and Atg14L seems to divert hVps/class III PI(3)K into an in perinuclear aggregates, where they co-localize with the aggresome autophagic role30, we investigated the interaction between Atg14L and marker HDAC6 (ref. 31) (Fig. 2c; Supplementary Information, Fig. S2d) beclin 1. Atg14L immunoreactivity was detected on beclin 1 immu- and also with γ-tubulin and are associated with the collapse of the vimen- noprecipitates and increased after starvation in both C38 and IB3-1 tin cage (data not shown)12,31. The absence of the beclin 1 interactome cells (Supplementary Information, Fig. S2a). Confocal microscopy and at the endoplasmic reticulum level18,19 might have functional relevance FRET analysis revealed co-localization of beclin 1 with hVps34, hVps 15, in autophagy deficiency of CF epithelia, because PI(3)K complex III is a Ambra 1 and Atg14L (Fig. 2a; Supplementary Information, Fig. S2c, d) specialized endoplasmic reticulum membrane platform for the assembly in both C38 and IB3-1 cells. hVps34 (Fig. 2b), beclin 1 and the other bec- of a phagosomal initiation complex18,19. lin 1-interacting proteins (data not shown) co-localized with the endo- Decreased protein levels of beclin 1, hVps34, hVps15, Ambra1 plasmic reticulum marker calnexin in C38 cells but not in IB3-1 cells. and Atg14L (Fig. 2d), and also of UVRAG (data not shown), were In CF cells, beclin 1 and beclin 1-interacting proteins are sequestered observed in IB3-1 cells. Western blots of the insoluble protein fraction

abHuman CF nasal mucosa Medium Cystamine cdHuman CF nasal mucosa CftrF508del mice Cystamine –+ – NAC 3-MA + NAC + cystamine + NAC Cystamine – + Mr(K) NAC –+– M (K) 3-MA + cystamine r Beclin 1 55 55 * Beclin 1 10 * 55 * 55 Tubulin 8 * Tubulin of tissue 2 6 + 3-MA 4 + cystamine + NAC 2 0 Human CF nasal mucosa LC3+ dots per mm CftrF508del mice e 10 PBS Cystamine * f g + cystamine 8 Cystamine –+ +

of tissue + PBS + PBS + 3-MA 2 3-MA –+– M (K) 6 r TG2 72 4 55 Tubulin SUMO-1 2 CftrF508del mice

F508del LC3+ dots per mm 0 Cftr mice +3-MA h i j k lm PBS PBS LV- LV- GFP beclin 1 TG2 activity LV-GFP LV-GFP Cystamine Cystamine Mr(K) 80 ) 12

LV-Beclin 1 –1 LV-Beclin 1

) TG2 72

–1 10 12 * of tissue 2

80 55 60 OD S of tissue * 10 –3 2 Tubulin 8 OD s –3 60 8 CftrF508del mice +LV-GFP 40 6 6 40 4 4 20 20 2

2 +LV-

beclin 1 L MPO activity (10 0 0 0 0 MPO activity (10 Macrophages per mm Macrophages +3-MA +3-MA CftrF508del mice CftrF508del mice CftrF508del mice Macrophages per mm Macrophages CftrF508del mice CftrF508del mice

Figure 7 Restoring autophagy by cystamine or beclin 1 overexpression different randomly taken sections. Means ± s.d. per mm2 of lung tissue rescues CF phenotype in human and mice CF airways. CF human from seven mice in each group. Asterisk, P < 0.01; ANOVA. (i) MPO nasal mucosae cultured with cystamine or NAC (patients 4–10 in activity in lung homogenates (expressed as 10–3 × optical density Supplementary Information, Table S1) (a–c) in the presence or absence (OD) s–1). Means ± s.d. for lung tissues from seven mice in each group. of 3-MA (patients 6–10 in Supplementary Information, Table S1) (b, c). Asterisk, P < 0.05; ANOVA. (j–m) CftrF508del mice (n = 3 per group) (a) Immunoblot of beclin 1. (b) Fluorescence microscopy quantification administered intranasally with a lentiviral vector encoding beclin 1 of LC3-positive dots. Results represent means ± s.d. Asterisk, P < 0.05 (LV-beclin 1) or GFP (LV-GFP). (j) TG2 Immunoblot in lung tissues. between groups of treatment; ANOVA. (c) Confocal microscopy Uncropped images of blots are shown in Supplementary Information, micrographs of p62. Nuclei counterstained with DAPI (blue). Scale Fig. S10. (k) Confocal microscopy of TG2 activity. Representative bar, 10 μm. Representative of five patients per group. (d–i) CftrF508del stainings from at least five airway epithelial areas per mouse from three mice treated with cystamine or PBS in the presence or absence of 3-MA mice per group. L, lumen. Scale bar, 10 μm. (l) Number of CD68+ (e–i) (n = 7 per group). (d) Immunoblot of beclin 1. (e) Quantification macrophages counted in 15–20 different randomly taken sections. of LC3 dots (per mm2 of lung tissue). Means ± s.d. Asterisk, P < 0.01; Means ± s.d. per mm2 of lung tissue. Asterisk, P < 0.01; ANOVA. ANOVA. (f) Immunoblot of TG2. Uncropped images of blots are shown (m) MPO activity in lung homogenates (expressed as 10–3 × OD s–1). in Supplementary Information, Fig. S10. (g) Confocal microscopy of Means ± s.d. for lung tissues. Asterisk, P < 0.05; ANOVA. Anti-β-Tubulin SUMO-1 in lung tissues. Nuclei counterstained with DAPI (blue). Scale was used as loading control for immunoblot analysis. Densitometric bar, 10 μm. (h) Number of CD68+ macrophages counted in 15–20 analysis of blots is shown in Supplementary Information, Fig. S9. revealed increased beclin 1 levels in IB3-1 cells in comparison with (Fig. 2f) and autophagosome formation (Fig. 2g), increased LC3 II levels C38 cells, suggesting the retention of beclin 1 as an insoluble pro- and decreased p62 accumulation (Fig. 2h, i). tein within aggregates in CF epithelia (Supplementary Information, Decreased beclin 1 protein levels were also observed in CF nasal Fig. S2e). polyp biopsies10,15,22 (n = 10) (Supplementary Information, Table S1 and Next we overexpressed beclin 1 in IB3-1 cells. Haemagglutinin Fig. S2f) and in airway tissues from CftrF508del mice (n = 10) (Fig. 2j), (HA)-tagged beclin 1 was detected at the endoplasmic reticulum level even though beclin 1 mRNA levels were similar to those in controls (Fig. 2e, top panel) and restored co-localization of hVps34 (Fig. 2e, bot- (data not shown). To investigate whether rescuing beclin 1 was effec- tom panel), hVps15, Ambra1 and Atg14L (data not shown) with cal- tive in restoring autophagy in vivo in CF airways, CftrF508del mice (n = 3 nexin. Moreover, HA–beclin induced an increase in LC3 dots (P < 0.001) per group) were administered intranasally with a lentiviral vector

abCftrF508del mice CftrF508del mice c CftrF508del mice d CftrF508del mice PBS NAC + PBS + PBS + PBS 10 + NAC + NAC + NAC 80

) 12

8 –1

of tissue 10

2 60 LC3 * 6 OD S –3 8 of tissue 2 40 6 * 4 * 4 2 20

per mm 2 p62 Number of macrophages LC3+ dots per mm

0 0 MPO activity (10 0

fg140 Scnn 1b-Tg mice 10 Scnn 1b-Tg mice e PBS NAC + PBS 130 8 + NAC Al546, TG2

of tissue Cy5, SUMO-1 2 120 6 *

LC3 110 ) Al546 (a.u.) 4 0

( F/F 100 WT + PBS 2 Scnn1b-Tg + PBS 90 Scnn1b-Tg + NAC LC3+ dots per mm 0 0204060 Time frame (s) ijScnn 1b-Tg mice Scnn 1b-Tg mice k

+ PBS + NAC h WT Scnn1b-Tg ) PBS PBS NAC ––+ 80 –1 12 Mr(K) 10 TG2 72 60 OD s –3 8 of tissue

2 40 p42/44 * 6 36 4 * 20 55 per mm 2 Tubulin 0 NAC Number of positve cells 0 NAC COX-2 MPO activity (10 Scnn 1b-Tg mice Scnn 1b-Tg mice

Histology COX-2

Figure 8 NAC restores autophagy and ameliorates CF lung phenotype in treated mice; ANOVA. (g) FRET analysis of TG2-Alexa 546 fluorescence after CF mice. CftrF508del mice (a–d) and Scnn1b-Tg mice (e–k) treated with NAC SUMO-1 Cy5 photobleaching in lung tissues from PBS-treated WT mice or PBS (n = 10 per group). (a) Confocal microscopy of LC3 (top) and p62 and PBS-treated or NAC-treated Scnn1b-Tg mice. (h) Immunoblot of TG2 (bottom) in lung tissues from CftrF508del mice. Nuclei counterstained with DAPI and phospho-p42/44 in lung tissues from PBS-treated WT mice and PBS- (blue). Scale bar, 10 μm. Representative images of ten PBS-treated and ten treated or NAC-treated Scnn1b-Tg mice. αβ-Tubulin was used as loading NAC-treated mice. (b) Quantification of LC3 dots (per mm2 of tissue) in lung control. (i) Lung histology of PBS-treated and NAC-treated Scnn1b-Tg mice. tissues from CftrF508del mice. Means ± s.d. Asterisk, P < 0.001 versus PBS- Haematoxylin staining. Scale bar, 150 μm. Black arrows indicate decrease treated mice; ANOVA. (c) Number of macrophages per mm2 of lung tissue in lung infiltration in NAC-treated in comparison with PBS-treated mice from CftrF508del mice (CD68-positive cells in 15–20 different randomly taken (arrowheads). (j) Left: confocal microscopy of COX-2 staining (left) in PBS- sections for each mouse lung for each condition); means ± s.d. for three treated and NAC-treated Scnn1b-Tg mice. Nuclei counterstained with DAPI separate experiments. Asterisk, P < 0.05 versus PBS-treated mice; ANOVA. (blue). Scale bar, 150 μm. Right; number of COX-2-positive cells per mm2 of (d) MPO activity in lung homogenates from CftrF508del mice (expressed as lung tissue in PBS-treated and NAC-treated Scnn1b-Tg mice. Means ± s.d. 10–3 × OD s–1). Means ± s.d. Asterisk, P < 0.05 versus PBS-treated mice; Asterisk, P < 0.05 versus PBS-treated mice; ANOVA. (k) MPO activity in lung ANOVA. (e) Confocal microscopy of LC3-positive dots in lung tissues from homogenates from PBS-treated and NAC-treated Scnn1b-Tg mice (expressed PBS-treated or NAC-treated Scnn1b-Tg mice. Nuclei counterstained with as 10–3 × OD s–1). Means ± s.d. Asterisk, P < 0.05 versus PBS-treated DAPI (blue). Scale bar, 10 μm. (f) Number of LC3-positive dots in lung mice; ANOVA. Densitometric analysis of the blot is shown in Supplementary tissues from Scnn1b-Tg mice. Means ± s.d. Asterisk, P < 0.01 versus PBS- Information, Fig. S9. encoding beclin 1 under the control of the ubiquitous CMV pro- TG2-mediated crosslinking induces sequestration of beclin 1 moter32,33 or a lentiviral vector expressing GFP under the control of in aggresomes and drives defective autophagy in CF airway the same promoter (LV-GFP) (Supplementary Information, Fig. S2g). epithelial cells Overexpression of beclin 1 from the lentivirus induced a significant These findings suggest that there is post-transcriptional regulation increase in LC3 dots (P < 0.001) and decreased p62 accumulation in of beclin 1 expression in CF airways. Because the beclin 1 protein comparison with CftrF508del mice treated with LV-GFP (Fig. 2k, l). sequence contains QP and QXXP motifs, which are specific target sites

NATURE CELL BIOLOGY ADVANCE ONLINE PUBLICATION 9 © 2010 Macmillan Publishers Limited. All rights reserved. ARTICLES for TG2 activity34 (Supplementary Information, Fig. S3a), and TG2 is (Fig. 4b). These effects were prevented by cystamine and by EUK-134 an autophagy inhibitor in pancreatic adenocarcinoma cells35, we tested (P < 0.01) (Fig. 4a, b) or NAC (data not shown). Transfection of C38 cells the hypothesis that beclin 1 might undergo TG2-mediated crosslinking. with HA–beclin on CFTR inhibition abrogated the effects on autophagy – Immunoprecipitates of beclin 1 from IB3-1 cells were blotted with an of CFTRinh-172 (data not shown). 16HBE14o cells behaved similarly to Nε(γ-l-glutamyl)-l-lysine isopeptide crosslink catalysed by TG2 (ref. 34). C38 cells (data not shown). F508del We observed high-molecular-mass bands in the Mr 130,000–250,000 Overexpression of GFP–CFTR in A549 lung adenocarcinoma cell (130K–250K) range that were decreased by knockdown of TG2 as well lines31 increased intracellular ROS and Ca2+ concentrations and eIF-2α as by the TG2 inhibitor cystamine34 (Fig. 3a; Supplementary Information, phosphorylation (Supplementary Information, Fig. S4b–d), thus indicat- Fig. S3b) or the calcium chelator bis-(o-aminophenoxy)ethane-N,N,Nʹ,Nʹ- ing CFTRF508del-induced activation of the unfolded protein response and tetra-acetic acid acetoxymethyl ester (BAPTA-AM)10,15 (data not shown) endoplasmic reticulum stress36,37. GFP–CFTRF508del also increased PIASy and enhanced by the proteasome inhibitor MG132 (Fig. 3b). and TG2 levels (Fig. 4c) and TG2 SUMOylation (Fig. 4d), and decreased TG2 short interfering RNA (siRNA) or cystamine decreased bec- beclin 1 and LC3 II protein levels (Fig. 4e) and LC3 dots (P < 0.001) lin1/HDAC6 co-localization and aggresome sequestration (Fig. 3c, d), (Fig. 4f). EUK-134 and cystamine (Fig. 4d–f) or NAC (data not shown) restored the soluble forms of beclin 1 and beclin 1 interactor proteins prevented these effects. Overexpression of mutant CFTR also induced (Fig. 3e), decreased the amount of beclin 1 in the insoluble protein frac- p62 accumulation and co-localization of p62 with GFP–CFTRF508del in tion (Supplementary Information, Fig. S3c) and increased the number of aggregates (Fig. 4g). These results indicate that defective CFTR has a Atg16L-positive puncta (Supplementary Information, Fig. S3d), which pivotal role in driving beclin 1 downregulation and defective autophagy associate transiently with the surface of forming autophagosomes24,36. in CF airways through the ROS–TG2 pathway. They also increased LC3 dots (P < 0.001) and autophagosome formation (Fig. 3g, h; Supplementary Information, Fig. S3e) and decreased the Restoring beclin 1 and autophagy decreases aggresome accumulation of p62 in IB3-1 cells (Fig. 3i). To prove the role of TG2 accumulation in CF epithelia further, we transfected autophagy-competent C38 cells with pLPCX-TG2, Next we investigated whether restoring beclin 1 levels might influ- a plasmid expressing wild-type TG2 or mutant Cys277Ser-TG2 including ence the aggresome-prone behaviour of CF epithelia. HA–beclin 1 the amino-acid substitution Cys277Ser in the catalytic site. Wild-type decreased the aggresome sequestration of PPAR-γ (Fig. 5a; TG2 induced a decrease in beclin 1 protein in C38 cells, whereas mutant Supplementary Information, Fig. S5a) and allowed PPAR-γ to interact TG2 did not (Fig. 3f). Overexpression of mutant TG2 restored beclin 1 with the nuclear co-repressor (N-CoR)15 in response to the PPAR-γ protein levels in IB3-1 cells (Fig. 3f). agonist rosiglitazone (Fig. 5b). Both HA–beclin 1 and cystamine These results indicate that TG2-mediated crosslinking drives defective induced CFTR co-localization with LC3 (Fig. 5c) and with the lyso- autophagy in CF epithelial cells. somal marker LAMP-1 (Fig. 5d). They also decreased CFTR/HDAC-6 co-localization and 3-MA (3-methyl adenine) prevented the effects ROS-mediated TG2 SUMOylation is responsible for defective of cystamine (Fig. 5d; Supplementary Information, Fig. S5b). Both autophagy in CF epithelia beclin 1 knockdown and 3-MA also inhibited the effects of cystamine Because increased ROS sustain high TG2 levels through TG2 in decreasing CFTR (Fig. 5e) and PPAR-γ (data not shown) seques- SUMOylation mediated by PIASy (protein inhibitor of activated STAT tration in aggresomes and co-localization with p62 (Fig. 5e). These y)15, we incubated IB3-1 cells with superoxide dismutase (SOD)–catalase results suggest that the effects of cystamine are due to its ability to mimetic EUK-134 or knocked-down PIASy15. We observed decreased rescue beclin 1 and autophagy. beclin 1 crosslinking and aggresome sequestration (data not shown), We investigated the role of p62 in driving aggresome formation in CF increased LC3 dots (P < 0.001) (Fig. 3g), autophagosome formation epithelia. p62 is a regulator of packing and targeting polyubiquitylated pro- (Fig. 3h) and beclin 1 upregulation (Fig. 3j); overexpression of man- teins and aggregates to autophagosomes for lysosomal degradation3,4 and ganese superoxide dismutase (MnSOD)15 and the ROS scavenger is involved in the formation of protein aggregates in cancer cells38,39. p62 N-acetylcysteine (NAC) showed similar effects to those of EUK-134 knockdown in autophagy-deficient IB3-1 cells decreased PPAR-γ and the (data not shown). NAC, EUK-134 or PIASy knockdown were all as effec- inhibitory protein IkB-α aggresomes (data not shown), restored the interac- tive as TG2 knockdown in restoring protein levels of soluble beclin 1, tion between PPAR-γ and N-CoR15 (Fig. 5b), decreased the accumulation of Atg14L, hVps34, hVps15 and Ambra1 (data not shown). These results CFTR aggregates (Supplementary Information, Fig. S5c) and decreased the indicate that ROS-mediated TG2 SUMOylation is responsible for defec- co-localization of CFTR with HDAC6 (data not shown). This suggests that, tive autophagy in CF epithelia. in CF epithelia, p62 mediates the effects of defective autophagy in driving aggresome accumulation of misfolded or modified proteins. Defective CFTR inhibits autophagy through ROS-mediated TG2 Defective autophagy with p62 accumulation compromises the delivery SUMOylation of ubiquitylated substrates to the proteasome40. However, upregulation of We tested whether defective autophagy is a consequence of the CF autophagy has been reported as a compensatory response to proteasome genetic defect, because ROS-mediated TG2 SUMOylation is induced on inhibition40, thus revealing crosstalk between the proteasome-based and inhibition of CFTR10,15. CFTR knockdown (Supplementary Information, the autophagy-based degradasomes40. We incubated IB3-1 cells with Fig. S4a) in autophagy-competent C38 cells decreased beclin 1 and LC3 II the proteasome inhibitor MG132 and demonstrated that MG132 was protein levels (Fig. 4a), LC3 dots (P < 0.001) (Fig. 4b) and increased p62 not effective in enhancing soluble beclin 1 protein or increasing LC3 levels (Fig. 4a). Similar effects were observed on inhibition of CFTR dots, but enhanced beclin 1 sequestration in HDAC6-positive aggregates 10,15 by the small molecule CFTR inhibitor-172 (CFTRinh-172) (P < 0.001) (Supplementary Information, Fig. S5d–f).

Rescuing beclin 1 and autophagy restores CFTR trafficking in decreased p42/44 phosphorylation (Fig. 6f) and TG2 SUMOylation CF epithelia (Supplementary Information, Fig. S6f) and increased the levels of Mr 55K These results indicate that HA–beclin 1, cystamine and p62 knockdown PPAR-γ protein (data not shown). In contrast, MG132 did not decrease all favoured the clearance of CFTRF508del aggregates. They also raise the TNF-α secretion in medium (P < 0.01) (Supplementary Information, question of whether rescuing beclin 1 and autophagy in IB3-1 cells might Fig. S5g) or p42/44 phosphorylation (data not shown). restore mutant CFTR maturation and trafficking to the Golgi and ulti- Because targeting TG2 restores inflammation in human and mice mately to the cellular surface. CF airways15, we investigated whether these effects were mediated by The F508del mutation results in a temperature-sensitive folding defect the rescue of autophagy. We incubated IB3-1 cells with cystamine or of CFTR and premature degradation by the endoplasmic-reticulum-asso- the antioxidants EUK-134 and NAC in the presence or absence of bec- ciated degradation system (ERAD)41,42. Functional CFTR can be increased lin 1 knockdown. The effects of cystamine or antioxidants (data not by growth at a lower temperature or incubation with chemical chaper- shown) in controlling p42/44 phosphorylation (Fig. 6g) and TNF-α ones, which favour the escape of CFTRF508del from ERAD41–47. secretion (Fig. 6h) were neutralized on beclin 1 knockdown (Fig. 6g, h; We transfected autophagy-deficient IB3-1 cells for 24 h with GFP– Supplementary Information, Fig. S6g). The effects of cystamine on CFTRF508del in the presence or absence of cystamine at 37 °C or p62 knock- TG2 SUMOylation and inflammation were also neutralized by either down. Either cystamine or p62 knockdown (data not shown) induced 3-MA48 (data not shown) or ATG5 knockdown (Fig. 6i, Supplementary an increase in GFP–CFTRF508del co-localization with the Golgi marker Information, Fig. S6h). GM-130 (Fig. 5f) and decreased the sequestration of CFTR in aggre- This indicates that rescue of beclin 1 and autophagy mediates the somes (Fig. 5g). Western blotting of the whole protein extracts revealed effects of cystamine and antioxidant molecules in CF epithelia. increased mature complex-glycosylated CFTRF508del (band C, apparent 46 Targeting the ROS–TG2 pathway ameliorates the CF phenotype Mr about 170K) after treatment with cystamine or p62 knockdown in comparison with cells transfected with GFP–CFTRF508del alone (Fig. 6a), through restoration of autophagy in human and mouse CF airways with an increased ratio between the complex-glycosylated band C and To validate these results in human CF airways, we cultured nasal 46 its core-glycosylated counterpart band B (apparent Mr about 150K) , polyp biopsies with cystamine or EUK-134 (refs 10, 15, 22). In all CF the predominant form in untreated cells46 (Fig. 6a). These results also biopsies tested (n = 7) (Supplementary Information, patients #4-10 revealed a greater increase in band B in cystamine-treated cells or after of Supplementary Information, Table S1), both NAC and cystamine p62 knockdown (Fig. 6a). Western blots of the insoluble protein frac- increased beclin 1 protein and LC3 dots (P < 0.05), decreased p62 tion revealed higher amounts of CFTRF508del band B that were decreased aggregates (Fig. 7a–c, top panel, Supplementary Information, Fig. S7a), after treatment with cystamine or p62 knockdown (Supplementary decreased TG2 SUMOylation and enhanced PPAR-γ/N-CoR interac- Information, Fig. S6a), indicating the sequestration of CFTRF508del within tion; these effects were prevented by 3-MA48 (Fig. 7b, c; Supplementary insoluble aggregates. Information, Fig. S7b, c). Confocal microscopy revealed an increase in GFP–CFTRF508del at the To confirm the biological relevance of these findings in vivo, we plasma membrane after treatment with cystamine (Fig. 6b, c) as well as treated CftrF508del homozygous mice with cystamine15 in the presence or on p62 knockdown (data not shown) in comparison with IB3-1 cells absence of 3-MA or PBS (n = 7 for each group). Cystamine has been transfected with GFP–CFTRF508del alone (Fig. 6b, c). To confirm the pres- proved to be effective in vivo in a mouse model of Huntington’s dis- ence of complex-glycosylated CFTR at the cell surface, plasma membrane ease49, and also to control inflammation in CftrF508del homozygous mice fractions were isolated from GFP–CFTRF508del-transfected IB3-1 cells by 15,23. In all mice tested, cystamine increased beclin 1 protein (Fig. 7d) biotinylation assay and streptavidin pull-down, as reported previously47. and LC3 dots (P < 0.01) (Fig. 7e), thus indicating rescue of autophagy. Band C of GFP–CFTRF508del was pulled down specifically on streptavidin These effects were neutralized in lung homogenates by 3-MA (Fig. 7e), beads after cystamine treatment (Fig. 6d, Supplementary Information, which also prevented the effects of cystamine in reducing TG2 protein Fig. S6b) or p62 knockdown (data not shown). (Fig. 7f), SUMO-1 protein expression15 (Fig. 7g), TG2 SUMOylation These results indicate that restoring intracellular homeostasis by res- (Supplementary Information, Fig. S7d), macrophage lung infiltration cuing beclin 1 and autophagy by cystamine or reducing p62 accumula- (per mm2 of tissue, versus PBS-treated mice; P < 0.01) (Fig. 7h) and mye- tion inhibits sequestration of mutant CFTR in aggresomes and allows loperoxidase (MPO) activity, a marker of neutrophilic recruitment and its maturation and trafficking to the cell membrane. activation status23 (Fig. 7i). On its own, 3-MA did not induce changes in lung phenotype in CftrF508del homozygous mice (data not shown). Restoring beclin 1 rescues the CF inflammatory phenotype To investigate whether beclin 1 overexpression was effective in damping Next we investigated whether defective autophagy might have a function down CF lung inflammation in vivo, we evaluated inflammation markers in inflammation in CF. HA–beclin 1 rescued pro-inflammatory phe- in LV-beclin 1-treated CftrF508del homozygous mice. Intranasal adminis- notype of IB3-1 cells as it decreased p42/44 phosphorylation (Fig. 6e), tration of LV-beclin 1 was effective in restoring autophagy in CftrF508del increased the levels of Mr 55K PPAR-γ protein (Fig. 6e) and decreased homozygous mice, as shown in Fig. 2. We demonstrate that LV-mediated tumour necrosis factor (TNF)-α secretion in medium (P < 0.01) beclin 1 overexpression induced a decrease in TG2 protein and activity (Supplementary Information, Fig. S6c). We also investigated whether (Fig. 7j, k) and decreased macrophage lung infiltration and MPO activ- defective autophagy might in turn sustain TG2 activation. HA–beclin 1 ity in lung homogenates (Fig. 7l, m), in comparison with LV-GFP-treated decreased ROS (P < 0.01) and TG2 SUMOylation (Supplementary mice (Fig. 7j–m). Information, Fig. S6d, e), and also PIASy and TG2 protein levels These results indicate that targeting ROS–TG2 axis ameliorates the human (Fig. 6e). Moreover, p62 knockdown in autophagy-deficient IB3-1 cells and mouse CF airway phenotype by restoring beclin 1 and autophagy.

Antioxidant treatment restores autophagy and ameliorates the off the post-translational regulatory mechanisms. As a consequence of CF lung phenotype in mouse CF airways this function, TG2 is involved in the pathogenesis of neurodegenerative To study the effects of NAC in vivo, we treated CftrF508del mice with daily diseases due to protein aggregates60. intraperitoneal injections of NAC50 (n = 10) or PBS (n = 10). NAC-treated Our study highlights an unpredicted role of p62 in CF. p62 is a mice showed an increase in LC3-positive vesicles (Fig. 8a) and punctate stress-induced protein that also regulates the formation and disposal dots (P < 0.001) (Fig. 8b), and a decrease in p62 accumulation (Fig. 8a), of the intracellular aggregates3,4,5,38–40. Either restoring beclin 1 by macrophage infiltration (P < 0.05) (Fig. 8c) and MPO activity (P < 0.05) cystamine or reducing p62 levels allows CFTRF508del maturation and in the lungs (Fig. 8d). trafficking to the cell surface, probably interrupting the cycle that We tested the effects of NAC in Scnn1b-Tg mice, another well-estab- disregulates the intracellular chaperone system. lished mouse model of CF, with overexpression of the β-ENaC subunit51 CF airway tissues are characterized by a pro-inflammatory phenotype and Na+ hyperabsorption, which induce a severe CF-like lung disease with constitutive ceramide accumulation61 and activation of NF-κB second- including inflammation51,52. Similarly to the Cftrf508del mice, Scnn1b-Tg ary to defective CFTR function62. We have previously reported that CFTR show a spontaneous lung inflammatory phenotype regardless of the inhibition results in the upregulation of ROS and their downstream events, presence of bacterial infections15,23,51,52. Scnn1b-Tg lung tissues showed leading to inflammation10,15. Here we show that CFTR knockdown drives decreased LC3-positive vesicles and LC3 dots and p62 accumulation, defective autophagy, as it occurs in human and mouse epithelia that are and increased TG2 activity, in comparison with their control littermates either homozygotes for F508del or compound heterozygotes for two severe (Supplementary Information, Fig. S7e–g). NAC-treated Scnn1b-Tg mice CFTR mutations. Because under the latter conditions negligible amounts of (n = 10) showed increased LC3-positive vesicles and LC3 dots (P < 0.01) functional CFTR are detected at the cell surface45,63,64, our results highlight (Fig. 8e, f) with decreased p62 aggregates (data not shown) and a decrease an unforeseen role of CFTR in the regulation of autophagy in CF. in TG2 SUMOylation (Fig. 8g), TG2 protein levels and p42/44 phospho- We show that targeting TG2 SUMOylation with antioxidant molecules rylation (Fig. 8h) in comparison with PBS-treated mice. Treatment with as well as inhibiting TG2 with cystamine rescue the CF lung phenotype NAC also decreased lung infiltration by leukocytes (Fig. 8i), decreased the by means of autophagy induction in both human and mouse CF airways. number of cyclo-oxygenase-1 (COX-2)-positive cells (per mm2 of tissue, Our results provide a new mechanism linking the CFTR defect with P < 0.05 versus PBS-treated mice) (Fig. 8j) and decreased MPO activity in inflammation, through the ROS–TG2-mediated inhibition of autophagy lung homogenates (P < 0.05 vs. PBS-treated mice) (Fig. 8k). (Supplementary Information, Fig. S8a, b). We have therefore provided a Finally, we found that cystamine increased LC3 dots and decreased novel rationale and a mechanism of action for both NAC65 and cystamine inflammatory lung infiltration (data not shown) in Scnn1b-Tg mice. Neither in the treatment of patients with CF. cystamine nor NAC induced changes in wild-type mice (data not shown). Our study suggests that the restoration of beclin 1 and autophagy may These results indicate that NAC rescues beclin 1 and autophagy and be a novel approach to the treatment of CF and may pave the way for the ameliorates the airway phenotype in vivo in CF mouse models. development of a new class of drugs that, by enhancing beclin 1 levels, could be effective treatments for CF. DISCUSSION Defective autophagy is a critical mechanism in several chronic human METHODS diseases53 such as neurodegeneration54 and cancer55. Here we have shown Methods and any associated references are available in the online version that defective autophagy due to decreased levels of beclin 1 protein, a of the paper at http://www.nature.com/naturecellbiology/ key molecule involved in autophagosome formation17,19,24,25,30, drives lung Note: Supplementary Information is available on the Nature Cell Biology website. inflammation in CF. Decreased beclin 1 expression has been described in the affected brain ACKNOWLEDGEMENTS regions in early Alzheimer’s disease, and defective autophagy has been We thank Noboru Mizushima for the gift of the pEGFP–LC3 and pcDNA3- 56 HA–beclin 1 expression vectors; Ron Kopito for the gift of the pGFP–F508del- linked to amyloid-β accumulation . A feature that is common to this neu- CFTR expression vector; Michael Bownlee for the gift of the adenoviral vectors; rodegenerative disease and CF is the aberrant intracellular accumulation of Gian Maria Fimia for the gift of the TG2 plasmid; Dieter C. Gruenert for the – – misfolded proteins within the affected tissues57,58. In CF airways, misfolded gift of CFBE41o and 16HBE14o cell lines; Maria Carla Panzeri for support in 31 10 electron microscopy and in the analysis of the data; Rosarita Tatè for technical or damaged proteins such as misfolded CFTR and PPAR-γ accumulate support in confocal microscopy; and Ilaria Russo for technical support in histology. in aggresomes. Here we demonstrate that in CF epithelia TG2-mediated Cftrtm1EUR (F508del (FVB/129) mice were obtained from Bob Scholte under crosslinking and sequestration of beclin 1 dislodge the PI(3)K platform European Economic Community European Coordination Action for Research in Cystic Fibrosis program EU FP6 LSHM-CT-2005-018932. This work was from the endoplasmic reticulum, thus inhibiting the initiation of autophagy. supported by the European Institute for Research in Cystic Fibrosis, Cancer TG2-driven defective autophagy in turn increases ROS and TG2 levels, thus Research UK, Rothschild Trust, Coeliac UK and Regione Campania (L. 229/99). generating a cycle leading to a deleterious pro-oxidative and pro-inflamma- AUTHOR CONTRIBUTIONS tory environment (Supplementary Information, Fig. S8a, b). A.L. co-designed the research concept, planned the overall experimental Autophagy inhibition with p62 accumulation3,4,5,38,39 increases the lev- design, performed organ culture and confocal microscopy studies and wrote els of proteasome substrates3 and compromises the ubiquitin proteasome the manuscript. V.R.V. co-designed the research concept, planned the overall 3,59 4,5 experimental design and performed immunoblot and immunoprecipitation system , thus favouring aggresome formation . TG2 contributes to experiments, cell cultures and transfections. S.E. contributed to the study proteasome overload and aggresome formation in CF airways by induc- design, interpretation and analysis of the data and performed immunoblot and ing the crosslinking and aggregation of several substrate proteins10,15. TG2 immunoprecipitation experiments, cell cultures and transfections. N.B. contributed to the study design, provided scientific knowledge, contributed to the interpretation therefore functions as a rheostat of the post-translational network and and analysis of the data, performed experiments on mice and wrote the manuscript. the ubiquitin proteasome system under disease conditions and switches D.M. contributed to the study design, provided scientific knowledge, contributed

to the interpretation and analysis of the data and performed the analysis of 30. Liang, C. et al. Beclin1-binding UVRAG targets the class C Vps complex to coordinate mitochondrial function. C.S. provided expression vectors and scientific knowledge autophagosome maturation and endocytic trafficking. Nature Cell Biol. 10, 776–787 and contributed to the analysis of the data. M.G. and L.P. performed experiments on (2008). mice and contributed to the interpretation and analysis of the data. I.G., M.P.M. and 31. Kawaguchi, Y. et al. The deacetylase HDAC6 regulates aggresome formation and cell viability in response to misfolded protein stress. Cell 115, 727–738 (2003). M.D. performed PCR and contributed to the interpretation and analysis of the data. 32. Spencer, B. et al. Beclin 1 gene transfer activates autophagy and ameliorates the neu- S.G. contributed to the discussion of the data. E.M. and B.S. provided the lentiviral rodegenerative pathology in α-synuclein models of Parkinson’s and Lewy body diseases. vectors and scientific knowledge. S.Q. contributed to the interpretation and analysis J. Neurosci. 29, 13578–13588 (2009). of the data and provided scientific knowledge. A.B. co-designed the research concept 33. Stocker, A. G. et al. Single-dose lentiviral gene transfer for lifetime airway gene expres- and co-supervised the project. V.R. and L.M. designed the research concept, planned sion. J. Gene Med. 11, 861–867 (2009). the overall experimental design, supervised the study and wrote the manuscript. 34. Lorand, L. & Graham, R. M. Transglutaminases: crosslinking enzymes with pleiotropic functions. Nature Rev. Mol. Cell Biol. 4, 140–156 (2003). COMPETING FINANCIAL INTERESTS 35. Akar, U. et al. Tissue transglutaminase inhibits autophagy in pancreatic cancer cells. Mol. Cancer Res. 5, 241–249 (2007). The authors declare no competing financial interests. 36. Ron, D. & Walter, P. Signal integration in the endoplasmatic reticulum unfolded protein response. Nature Rev. Mol. Cell Biol. 8, 519–529 (2007). Published online at http://www.nature.com/naturecellbiology 37. Bartoszewski, R. et al. Activation of the unfolded protein response by ΔF508 CFTR. Reprints and permissions information is available online at http://npg.nature.com/ Am. J. Respir. Cell Mol. Biol. 39, 448–457 (2008). reprintsandpermissions/ 38. Moscat, J. & Diaz-Meco, M. T. p62 at the crossroads of autophagy, apoptosis, and cancer. Cell 137, 1001–1004 (2009). 39. Mathew, R. et al. Autophagy suppresses tumorigenesis through elimination of p62. Cell 1. Mizushima, N., Levine, B., Cuervo, A. M. & Klionsky, D. J. Autophagy fights disease 137, 1062–1075 (2009). through cellular self-digestion. Nature 28, 1069–1075 (2008). 40. Komatsu, M. et al. Homeostatic levels of p62 control cytoplasmic inclusion body 2. Moreau, K., Luo, S. & Rubinsztein, D. C. Cytoprotective roles for autophagy. Curr. Opin. formation in autophagy-deficient mice. Cell 131, 1149–1163 (2007). Cell Biol. 22, 206–211 (2010). 41. Riordan, J. R. CFTR function and prospects for therapy. Annu. Rev. Biochem. 77, 3. Korolchuk, V. I., Mansilla, A., Menzies, F. M. & Rubinsztein, D. C. Autophagy inhibition 701–726 (2008). compromises degradation of ubiquitin–proteasome pathway substrates. Mol. Cell 33, 42. Wang, X. et al. Hsp90 cochaperone Aha1 downregulation rescues misfolding of CFTR 517–527 (2009). in cystic fibrosis. Cell 127, 803–815 (2006). 4. Kirkin, V., McEwan, D. G., Novak, I. & Dikic, I. A role for ubiquitin in selective 43. Wang, X., Koulov, A. V., Kellner, W. A., Riordan, J. R. & Balch, W. E. Chemical and 34, autophagy. Mol. Cell 259–269 (2009). biological folding contribute to temperature-sensitive ΔF508 CFTR trafficking. Traffic 5. Bjørkøy, G. et al. p62/SQSTM1 forms protein aggregates degraded by autophagy and has a 11, 1878–1893 (2008). protective effect on huntingtin-induced cell death. J. Cell Biol. 171, 603–614 (2005). 44. Skach, W. R. CFTR: new members join the fold. Cell 127, 673–675 (2006). 6. Dohm, C. P., Kermer, P. & Bahr, M. Aggregopathy in neurodegenerative diseases: mecha- 45. Amaral, M. D. CFTR and chaperones: processing and degradation. J. Mol. Neurosci. nisms and therapeutic implication. Neurodegen. Dis. 5, 321–338 (2008). 23, 41–48 (2004). 7. Williams, A. et al. Aggregate-Prone proteins are cleared from the cytosol by autophagy: 46. Pedemonte, N. et al. Small-molecule correctors of defective ΔF508-CFTR cellular Therapeutic Implications. Curr. Top. Dev. Biol. 76, 89–101 (2006). processing identified by high-throughput screening. J. Clin. Invest. 115, 2564–2571 8. Schessl, J., Zou, Y., McGrath, M. J., Cowling, B. S. & Maiti, B. Proteomic identification (2005). of FHL1 as the protein mutated in human reducing body myopathy. J. Clin. Invest. 118, 47. Caohuy, H., Jozwik, C. & Pollard, H. B. Rescue of ΔF508-CFTR by the SGK1/Nedd4-2 904–912 (2008). signaling pathway. J. Biol. Chem. 284, 25241–25253 (2009). 9. Rodriguez-Gonzalez, A. et al. Role of the aggresome pathway in cancer: targeting histone 48. Ghavami, S. et al. S100A8/A9 induces autophagy and apoptosis via ROS-mediated deacetylase 6-dependent protein degradation. Cancer Res. 68, 2557–2560 (2008). cross-talk between mitochondria and lysosomes that involves BNIP3. Cell Res. 20, 10. Maiuri, L. et al. Tissue transglutaminase activation modulates inflammation in cystic 314–331 (2010). fibrosis via PPARγ down-regulation. J. Immunol. 180, 7697–7705 (2008). 49. Karpuj, M. V. et al. Prolonged survival and decreased abnormal movements in transgenic 11. Ratjen, F. & Doring, G. Cystic fibrosis. Lancet 361, 681–689 (2003). model of Huntington disease, with administration of the transglutaminase inhibitor 12. Sha, Y., Pandit, L., Zeng, S. & Eissa, N. T. A critical role of CHIP in the aggresome cystamine. Nature Med. 8, 143–149 (2002). pathway. Mol. Cell. Biol. 29, 116–128 (2009). 50. Sablina, A. A., Budanov, A. V., Ilyinskaya, G. V., Agapova, L. S. & Kravchenko, J. E. 13. Bence, N. F., Sampat, R. M. & Kopito, R. R. Impairment of the ubiquitin–proteasome The antioxidant function of the p53 tumor suppressor. Nature Med. 11, 1306–1313 system by protein aggregation. Science 292, 1552–1555 (2001). (2005). 14. Fu, L. & Sztul, E. ER-associated complexes (ERACs) containing aggregated cystic 51. Mall, M., Grubb, B. R., Harkema, J. R., O’Neal, W. K. & Boucher, R. C. Increased airway fibrosis transmembrane conductance regulator (CFTR) are degraded by autophagy. epithelial Na+ absorption produces cystic fibrosis-like lung disease in mice. Nature Med. Eur. J. Cell Biol. 88, 215–226 (2009). 10, 15. Luciani, A. et al. SUMOylation of tissue transglutaminase as link between oxidative 487–493 (2004). 10, stress and inflammation. J. Immunol. 183, 2775–2784 (2009). 52. Frizzell, R. A. & Pilewski, J. M. Finally, mice with CF lung disease. Nature Med. 16. Trudel, S. et al. Peroxiredoxin 6 fails to limit phospholipid peroxidation in lung 452–454 (2004). from Cftr-knockout mice subjected to oxidative challenge. PLoS One 4, e6075 53. Maiuri, C., Zalckvar, E., Kimchi, A. & Kroemer, G. Self-eating and self-killing: crosstalk (2009). between autophagy and apoptosis. Nature Rev Mol. Cell. Biol. 8, 741–752 (2007). 17. Sinha, S. & Levine, B. The autophagy effector Beclin 1: a novel BH3-only protein. 54. Hara, T. et al. Suppression of basal autophagy in neural cells causes neurodegenerative Oncogene 27, S137–S148 (2008). disease in mice. Nature 441, 885–889 (2006). 18. Maiuri, M. C., Criollo, A. & Kroemer, G. Crosstalk between apoptosis and autophagy 55. Takahashi, Y. et al. Bif-1 interacts with Beclin 1 through UVRAG and regulates within the Beclin 1 interactome. EMBO J. 29, 515–516 (2010). autophagy and tumorigenesis. Nature Cell Biol. 9, 1142–1151 (2007). 19. He, C. & Levine, B. The Beclin 1 interactome. Curr. Opin. Cell Biol. 22, 140–149 56. Pickford, F. et al. The autophagy-related protein beclin 1 shows reduced expression (2010). in early Alzheimer disease and regulates amyloid beta accumulation in mice. J. Clin. 20. Klionsky, D. J., Abeliovich, H., Agostinis, P., Agrawal, D. K. & Aliev, G. Guidelines for Invest. 118, 2190–2199 (2008). the use and interpretation of assays for monitoring autophagy in higher eukaryotes. 57. Tebbenkamp, A. T. & Borchel, D. R. Protein aggregate characterization in models of Autophagy 4,151–175 (2008). neurodegenerative disease. Methods Mol. Biol. 566, 85–91 (2009). 21. Mizushima, N., Yoshimori, T. & Levine, B. Methods in mammalian autophagy research. 58. Martínez, A, Portero-Otin, M., Pamplona, R. & Ferrer, I. Protein targets of oxidative Cell 140, 313–326 (2010). damage in human neurodegenerative diseases with abnormal protein aggregates. Brain 22. Raia, V. et al. Inhibition of p38 mitogen activated protein kinase controls airway inflam- Pathol. 20, 281–297 (2010). mation in cystic fibrosis. Thorax 60, 773–780 (2005). 59. Korolchuk, V. I., Menzies, F. M. & Rubinsztein, D. C. Mechanisms of cross-talk between 23. Legssyer, R. et al. Azithromycin reduces spontaneous and induced inflammation in the ubiquitin–proteasome and autophagy–lysosome systems. FEBS Lett. 584, 1393– DF508 cystic fibrosis mice. Respir. Res. 7, 134 –136 (2006). 1398 (2010). 24. Matsunaga, K. et al. Two Beclin 1-binding proteins, Atg14L and Rubicon, reciprocally 60. Muma, N. A. Transglutaminase is linked to neurodegenerative disease. J. Neuropathol. regulate autophagy at different stages. Nature Cell Biol. 11, 385–396 (2009). Exp. Neurol. 66, 258–263 (2007). 25. Zhong, Y. et al. Distinct regulation of autophagic activity by Atg14L and Rubicon 61. Teichgräber, V. et al. Ceramide accumulation mediates inflammation, cell death and associated with Beclin 1-phosphatidylinositol-3-kinase complex. Nature Cell Biol. 11, infection susceptibility in cystic fibrosis. Nature Med. 14, 382–391 (2008). 468–476 (2009). 62. Vij, N., Mazur, S. & Zeitlin, P. L. CFTR is a negative regulator of NFκB mediated innate 26. Axe, E. L. et al. Autophagosome formation from membrane compartments enriched immune response. PLoS One 4, e4664 (2009). in phosphatidylinositol 3-phosphate and dynamically connected to the endoplasmic 63. Farinha, C. M. & Amaral, M. D. Most F508del-CFTR is targeted to degradation at an reticulum. J. Cell Biol. 182, 685–701 (2008). early folding checkpoint and independently of calnexin. Mol. Cell. Biol. 25, 5242–5252 27. Hayashi-Nishino, M. et al. A subdomain of the endoplasmic reticulum forms a cradle (2005). for autophagosome formation. Nature Cell Biol. 11, 1433–1437 (2009). 64. Scott-Ward, T. S. & Amaral, M. D. Deletion of Phe508 in the first nucleotide-binding domain of the cystic fibrosis transmembrane conductance regulator increases its affinity 28. Maiuri, M. C. et al. Functional and physical interaction between Bcl-XL and a BH3-like domain in Beclin-1. EMBO J. 26, 2527–2539 (2007). for the heat shock cognate 70 chaperone. FEBS J. 276, 7097–7109 (2009). 29. Pattingre, S. et al. Bcl-2 antiapoptotic proteins inhibit Beclin 1-dependent autophagy. 65. Rochat, T., Lacroix, J. S. & Jornot, L. N-acetylcysteine inhibits Na+ absorption across Cell 23, 927–939 (2005). human nasal epithelial cells. J. Cell Physiol. 201, 106–116 (2004).

METHODS Lentiviruses. The lentiviral vectors expressing mouse beclin 1 or GFP have been Cell lines. IB3-1 human CF bronchial epithelial, C38 or S9 cells, normal bronchial described previously33. epithelial 16HBE14o– or lung carcinoma A549 cell lines (LGC Promochem, Milan, Italy) were cultured as recommended by the American Type Culture Collection. Adenoviral vector. Human MnSOD cDNA was cloned into the shuttle vector CFBE41o– human CF bronchial epithelial cells (a gift from D. C. Gruenert) were pAd5CMVK-NpA (a gift from Michael Brownlee). Cell lines were infected for cultured in MEM Earl’s salt l-glutamine (200 mM l-glutamine) medium sup- 2 h with MnSOD or control adenovirus (ref. 15). plemented with 10% FBS and the appro priate amount of penicillin/streptomycin. Cell lines were incubated for 24 h with cystamine (250 μM; Sigma Aldrich), Quantitative RT–PCR. Total RNA was extracted with the RNeasy Mini Kit NAC (10 mM; Vinci Biochem), EUK-134 (50 μg ml–1; Vinci Biochem), rapamycin (Qiagen). The mRNA was reverse transcribed with a SuperScriptTM III First (100 nM; Sigma Aldrich), chloroquine (25 μM; Sigma Aldrich), E64d (50 μg ml–1 Strand Synthesis System (Invitrogen). Quantitative RT–PCR was performed with for 4 h) and pepstatin A (50 μg ml–1 for 4 h), rosiglitazone (10 μM for 4 h; Vinci an iCycler iQ Multicolour Real-Time PCR Detector (Bio-Rad) with iQ TM SYBR Biochem); BAPTA-AM (25 μg ml–1 for 4 h; Calbiochem) and proteasome inhibi- Green supermix (Bio-Rad). Expression levels of genes were normalized to β-actin tor MG132 (50 μM for 6 h; Calbiochem). IB3-1 cells were also incubated for 24 h levels in the same sample. The sequence of primers is reported in Supplementary with 3-MA (20 mM; Calbiochem) followed by cystamine. C38 and 16HBE cells Information, Table S2b. The relative amounts of mRNA were calculated by using 10,15 the comparative Ct method. were incubated for 48 h with CFTRinh-172 (20 μM; Calbiochem) followed by cystamine or EUK-134. Immunoblot analysis. Immunoblot analysis was performed as described Human subjects and ex vivo cultures of nasal polyp mucosal biopsies. Ten patients previously10,15. The antibodies used for immunoblot analysis are reported with CF carrying severe CFTR mutations (Supplementary Information, Table S1) in Supplementary Information, Table S3. Densitometric analysis was per- and ten consecutive patients without CF (mean age 21 years, range 16–32 years) formed with Image J software; each data point is expressed as a mean ± s.d. undergoing surgical treatment for non-allergic nasal polyposis were enrolled. Nasal for three independent experiments. The data are shown in Supplementary polyp biopsies from seven patients in each group (patients 4–10 in Supplementary Information. Information, Table S1) were cultured for 4 h with or without EUK-134 (50 μg ml–1), NAC (10 mM) or cystamine (400 μM). Nasal polyp biopsies from five patients with Immunoprecipitation. Whole-cell lysates were incubated at 4 °C for 8–12 h with CF (patients 6–10 in Supplementary Information, Table S1) were also incubated antibody anti-beclin 1 (Santa Cruz Biotechnology). After the addition of Protein with NAC or cystamine in the presence or absence of 3-MA (3 mM)20, followed by A/G-agarose beads, the incubation was continued for 1 h. After being washed, incubation wth rosiglitazone (10 mM for 4 h). the immunoprecipitated proteins were subjected to electrophoresis through 8% polyacrylamide gels (Bio-Rad), transferred to blotting membranes (Polyscreen Mice and treatments. Young adult female CF mice homozygous for the F508del poly(vinylidene difluoride); NEN) and analysed. mutation (abbreviated CftrF508del)23 in the 129/FVB outbred background (obtained from Bob Scholte, Erasmus Medical Center Rotterdam, The Netherlands), young Soluble and insoluble fractions. Cells were lysed in buffer containg 50 mM Tris- adult Scnn1b-Tg mice51 and their wild-type littermates were housed in static HCl pH 7.5, 150 mM NaCl, 0.5% Nonidet P40, 5 mM EDTA, 1 mM phenylmeth- isolator cages at the animal care specific pathogen-free facility of Charles River ylsulphonyl fluoride, 50 mM NaF, 10 μg ml–1 leupeptin and 10 μg ml–1 apoprotein (Varese, Italy). We treated CftrF508del homozygous mice and Scnn1b-Tg mice by supplemented with protease inhibitors (Sigma) and centrifuged at 16,000g at 4 °C intraperitoneal injection with a daily dose of 100 μl of 0.01 M cystamine in PBS for 20 min. After centrifugation the soluble (supernatant) and insoluble (pel- solution for 7 days (ref. 16) (n = 7 for each group of treatment) or NAC with a let) fraction were used in western blot analysis with anti-beclin 1 or anti-GFP daily dose of 100 mg per kg body weight in PBS solution for 5 days or with PBS antibody. The pellet insoluble in Nonidet P40 was dissolved five times in sample solution (n = 10 for each group of treatment). CftrF508del mice were also treated buffer, boiled at 95 °C for 5 min and resolved on a 10% polyacrylamide gel. with cystamine in the presence or absence of 3-MA (intraperitoneal injection of 50 mg kg–1 in 100 μl of PBS for 7 days) or 3-MA alone (n = 7 for each group of Cell-surface biotinylation assay. Cell-surface proteins were biotinylated with treatment). Airway administrations of lentiviral vectors expressing mouse bec- sulphosuccinimidyl-6-(biotinamido) hexanoate (sulpho-NHS-LC-biotin, 1 –1 47 lin 1 or GFP32 were performed in anaesthetized CftrF508del homozygous mice aged mg ml in PBS, pH 8.2; Pierce), as described . IB3-1 cells were homogenized in 4–8 weeks, as described previously33. In brief, mice received a single dose of 4 μl a Potter–Elvehjem pestle and glass tube (Sigma) and centrifuged at 2,300g for 15 of 1% l-α-lysophosphatidylcholine (LPC; Sigma) in PBS into the nasal airway by min at 4 °C to obtain nuclear pellets. Supernatants containing the cytoplasmic inhalation-driven instillation 1 h before delivery of the GFP or beclin 1 lentiviral fractions and plasma membrane were centrifuged for 1 h at 16,000g at 4 °C; the vectors at a dose of 1.2 × 107 transducing units per mouse delivered into the nose pellet was the intact membrane and was solubilized in buffer A containing 1% (two 10-μl aliquots over several minutes). Mice were killed 5 days after vector Triton X-100. This was then centrifuged for 1 h at 60,000g in the ultracentrifuge. administration, and lung tissues were collected for analyses. The supernatants were collected as the plasma membrane fraction. Equivalent amounts of protein (500 μg) were used for streptavidin–agarose pull-down Histology. We excised and immediately fixed mouse lungs in buffered 4% para- (Pierce). Biotinylated proteins were immunoblotted against GFP or E-cadherin formaldehyde; they were then embedded in paraffin and sectioned to a thickness or β-actin antibodies. Incubation without biotin was used as control. of 5 μm. We performed haematoxylin staining on serial sections. Immunofluorescence analysis. Immunofluorescence analysis was performed Plasmids. The pGFP–CFTRF508del expression vector (provided by Ron R. Kopito), as described previously10,15. Cell or tissue sections were incubated with pri- pEGFP–LC3 and pcDNA3-HA–beclin 1 expression vector (a gift from N. mary antibodies. The antibodies used for immunofluorescence analysis are Mizushima), pLPCX-TG and pLPCX-TGCys277Ser (a gift from G. M. Fimia) reported in Supplementary Information, Table S3. TG2 enzymatic activity in and the corresponding empty expression vectors were used for transfection human or mice lung sections was detected by incubating unfixed sections with experiments. biotin-mono-dansyl cadaverine for 1 h at room temperature, as described previ- ously10,15. Image J software (US National Institutes of Health) was used to quan- Transfection and RNA interference. Cells were transfected with TG2 (ref. 15), tify the number of GFP–LC3-positive cells containing more than five GFP–LC3 PIASy (ref. 15), beclin 1 and p62 (Oligonucleotide Synthesis Core Facility, CEINGE puncta or cells with more than LC3 dots per cell. At least 60 cells were counted s.c.a.r.l. Naples, Italy), CFTR and ATG5 siRNA (Invitrogen) or scrambled oligo- in each experiment and the values are means ± s.d. for five independent experi- nucleotides by using Lipo RnaiMax (Invitrogen) in accordance with the manufac- ments. Correlation coefficient analysis was used as value of co-localization as turer’s instructions. siRNA oligonucleotide sequences are shown in Supplementary described previously66. Information, Table S2a. Cell lines were also transfected with pcDNA3-HA–beclin 1 expression vectors, pEGFP–LC3, human TG2 or Cys277Ser TG2 or antisense FRET microscopy. For acceptor photobleaching, FRET microscopy analysis was cDNAs in pLPCLX vector and GFP–CFTRF508del, by using Lipofectamine 2000 performed as described previously10,15. The antibodies used for FRET microscopy (Invitrogen) in accordance with the manufacturer’s instructions. are reported in Supplementary Information, Table S3. Cy5 was bleached at not

14 NATURE CELL BIOLOGY ADVANCE ONLINE PUBLICATION © 2010 Macmillan Publishers Limited. All rights reserved. DOI: 10.1038/ncb2090 METHODS more than 10% of its initial fluorescence, by 200 2.56-ms pulses of 5-mW 633-nm Probes) and analysed with a Wallac 1420 multilabel counter as described laser per pixel, sampling 0.01 mm2 of the specimen. Alexa 546 fluorescence was previously10,15. detected before and after Cy5 photobleaching. Calcium imaging. We loaded A549 cells transfected with CFTRF508del or vector Electron microscopy. Cell culture monolayers were fixed for 15 min at 4 °C with control with Fura-2 (30 min at 37 °C) in calcium imaging buffer in accordance 4% paraformaldehyde and 2.5% glutaraldehyde in 125 mM cacodylate buffer and with the manufacturer’s instructions (Invitrogen, Molecular probes). stained with uranyl acetate and lead citrate and examined in a Leo912 electron microscope. We used Image J software (US National Institutes of Health) to quan- Enzyme-linked immunosorbent assay (ELISA). Human TNF-α secretion tify the number of autophagosomes per cell in 15–20 different randomly taken was measured with a BD OptEIATM TNF-α ELISA kit II (BD Biosciences, micrographs for each condition. NJ, USA). Values were normalized to 106 cells; results are expressed as means ± s.d. Mitochondrial membrane potential measurements. We incubated 106 PBS- –1 washed cells in 40 nM DiOC6 (Sigma) and 1 mg ml propidium iodide (Sigma) Statistical analysis. Data are reported as arithmetic means ± s.d. The data dis- for 15 min at 37 °C. After being washed, cells were suspended in 1 ml of PBS tribution was analysed for normality and statistical analysis was performed with (pH 7.4) and were subsequently analysed by flow cytometry. a one-way ANOVA. Significant differences are indicated in the figures. All data were obtained from independent measurements. Data were analysed with SPSS MPO activity. MPO activity in lung homogenates was assessed at 490 nm over 13 software. Statistical significance was defined as P < 0.05. 10 min and expressed as 10–3 × OD s–1, as described previously24. 66. Luciani, A. et al. Lysosomal accumulation of gliadin p31-43 peptide induces oxida- ROS detection. Cells were pulsed with 10 μM 5-(and-6)-chloromethyl-2ʹ,7ʹ- tive stress and tissue transglutaminase-mediated PPARγ downregulation in intestinal dichlorodihydrofluorescein diacetate acetyl ester (CM-H2DCFDA; Molecular epithelial cells and coeliac mucosa. Gut 59, 311–319 (2010).

DOI: 10.1038/ncb2090

Figure S1 Defective autophagy in human CF airway epithelia. (a) Electron 10 m. (d) Immunoblot of LC3 in C38 and IB3-1 cells in the presence or microscopy analysis of non-starved C38 and IB3-1 cells. Bar, 250 nm. (b) absence of rapamycin. β-actin was used as loading control. (e-f) Real-time Immunoblot of LC3 in starved IB3-1 cells in the presence or absence of PCR of ULK-1, ULK-2, ARTG5, ATG6, ATG7, ATG8, ATG14L, BCL-2, Vps34, chloroquine. β-actin was used as loading control. (c) Confocal microscopy Vps15 and Ambra-1 mRNA in C38 and IB3-1 cells (e) or 16HBE14 and micrographs of p62 in starved C38 and IB3-1 cells. N, nucleus. Scale bar, CFBE 41 (f).

Figure S2 Beclin 1 and beclin 1 interactor proteins in CF airways. (a,b ) IB3-1 fluorescence after hVps34-Cy5, hVps15-Cy5, Ambra-1-Cy5 and ATG14L-Cy5 and C38 cells were grown in nutrient-rich or starvation medium. (a, left) photobleaching. F/F0 indicates post-bleaching/pre-bleaching fluorescence Immunoprecipitated (IP) beclin 1 species from whole cell extracts from IB3-1 ratio of Al546/Cy5. (d) Confocal microscopy micrographs of IB3-1 cells of or C38 cells were subjected to immunoblotting with the antibodies against ATG14L (red) and beclin 1 (green) (left panel) or ATG14L (red) and HDAC-6 Ambra 1, hVps15, hVps34, Atg14L, Bcl-2, and beclin1. For full scan see (green) (right panel) antibodies. Yellow indicates co-localization. N, nucleus. Supplemtary Information, Fig.S10. (a, right) Densitometric analysis of the Scale bar, 10 μm. (e) Immunoblot of Beclin1 of soluble and insoluble detergent band intensity (mean ± s.d. of three independent experiments): fold increase (NP-40) fractions in starved C38 and IB3-1. (f) Beclin 1 immunoblot in control after starvation versus cells cultured in nutrient-rich conditions. (b) FRET and CF human nasal mucosa. αβ-tubulin was used as loading control. (g) GFP analysis of beclin 1-Alexa 546 fluorescence after Bcl-2-Cy5 photobleaching. and mouse beclin 1 expression in airway tissues from CftrF508del homozygous F/F0 indicates post-bleaching/pre-bleaching fluorescence ratio of Al546/ mice after intranasal administration of a lentiviral vector encoding beclin 1(LV- Cy5. (c) Starved IB3-1 and C38 cells. FRET analysis of beclin 1-Alexa 546 beclin1) or expressing the GFP (LV-GFP). L, Lumen. Scale bar, 10 μm.

Figure S3 Effects of TG on autophagy in airway epithelia. (a) Sites for TG2 ± s.d. of three independent experiments (*,p<0.01 vs cells transfected with activity on beclin 1 sequence. (b) Immunoprecipitated (IP) beclin 1 species control siRNA; ANOVA). (e) left panel, confocal microscopy micrographs of LC3 from whole cell extracts of IB3-1 is immunoreactive for the anti-isopeptide (green). DAPI (blue), nuclear counterstaining. Scale bar, 10 μm. right panel, antibody. Immunoprecipitated (IP) without antibody from whole cell extracts of percentage of cells containing more than five LC3 positive punctate dots. IB3-1 was used as control (c) Immunoblot analysis of beclin1 of soluble and Mean ± s.d. of three independent experiments (*, p<0.001 vs cells transfected insoluble detergent (NP-40) fractions in starved IB3-1 in presence or asbsence control siRNA; ANOVA). (f) left panel, electron micrographs of starved IB3-1 of cystamine. (d-f) IB3-1 cells were transfected with either TG2 siRNA or cells showing areas enriched in autophagosomes (arrow) upon TG2 siRNA. scrambled oligonucleotides upon starvation. (d) left panel, confocal microscopy Scale bar, 100 nm. righ panel, number of autophagosomes per cell, counted micrographs Atg16L (red). DAPI (blue), nuclear counterstaining. Scale bar, in 20 cells per experiment. Mean ± s.d. of three independent experiments. (*, 10 μm. right panel, quantification of the number of Atg16L dots/cell. Mean p<0.001 vs siRNA control cells; ANOVA).

H I

"555I"

"5 H

" α I

Figure S4 GFP-F508del-CFTR increases ROS and Calcium levels and activates chloromethyl-2’7’-dichlorodihydrofluorescein diacetate acetyl ester (CM- UPR in A549 cells. (a). Immunoblot analysis of CFTR in C38 cells transfected H2DCFDA). (c) Ca2+ imaging of GFP-F508del CFTR transfected A549 cells with either human CFTR siRNA or scrambled oligonucleotides. β-actin was loaded with FURA-2. Pseudocolor ratiometric images. Colors correspond to the 2+ 2+ used as loading control. (b-d ) A549 cells were transfected with 1 g of pGFP- scale of [Ca ]i increase. Red, high [Ca ]i contents. Quantification of results F508del-CFTR, or pGFP. (b) Increase of intracellular ROS in GFP-F508del from three indipendent experiments is shown in the histogram. *, p<0.001 vs CFTR transfected A549 cells (each bar represents the mean±s.d. of three empty vector. N, nuclei. (d) Immunoblot analysis of p-eIF-2 in GFP-F508del indipendent experiments). *, p<0.001 vs empty vector. DCF, 5-(and-6)- CFTR transfected A549 cells. β-actin was used as loading control.

4 4Q 8 Q 4 " Q@ Q @ Q "4 F γ

@ 4 @ Q "4

Q 4 4 4" " Q Q4 4F" 2

4 4

Q4 "

4F" Q QD 4

4F" " α

Figure S5 Beclin 1 overexpression and p62 knock down rescue aggresome counterstaining. Scale bar, 10 m. (d) Starved IB3-1 cells cultured with formation. (a) IB3-1 cells transfected with either HA-tagged human MG132. top panel, immunoblot of Beclin1. bottom panel, densitometric beclin 1 or empty vector or with p62 siRNA or scrambled oligonucleotides analysis normalized to β-actin (mean ± s.d. of three independent followed by rosiglitazone. Confocal microscopy analysis of PPAR (green). experiments). (e ) Confocal microscopy analysis of CFTR (green) and DAPI (blue), nuclear counterstaining. Scale bar, 10 m (b) IB3-1 HDAC-6 (red). Yellow colour indicates co-localization. N, nucleus. Scale cells transfected with HA-beclin 1, or empty vector, or with cystamine bar, 10 m. (f ) IB3-1 cells transfected with plasmid expressing GFP-LC3. in presence or absence of 3-MA. FRET analysis of CFTR-Alexa 546 left panel, confocal laser microscopy. Scale bar, 10 m. right panel, fluorescence after HDAC6 -Cy5 photobleaching. F/F0 indicates post- percentage of cells containing more than five GFP-LC3 punctate dots per bleaching/pre-bleaching fluorescence ratio of Al546/Cy5. (c) IB3-1 cells cell. Mean ± s.d. of five independent experiments. (g) TNF- secretion in transfected with 50 nM human p62 siRNA or scrambled oligonucleotudes. IB3-1 cells after MG132 treatment. Each bar represents the mean ± s.d. of Confocal microscopy analysis CFTR (green). DAPI (blue), nuclear three independent experiments. *, p<0.01 vs control ANOVA.

) α

Figure S6 Cystamine and p62 knock down restores F508del-CFTR trafficking the mean ± s.d. of three independent experiments.*, p<0.01 compared to and inflammatory phenotype in CF airway epithelia(a) IB3-1 cells were the empty vector (e) IB3-1 cells transfected with either human p62 siRNA or transfected with 1μg of pGFP-F508del-CFTR in presence of either human scrambled oligonucleotides. FRET analysis of TG2-Alexa 546 fluorescence p-62 siRNA or scrambled oligonucleotides. The cells were lyzed with detergent after SUMO-1-Cy5 photobleaching. F/F0 indicates post-bleaching/pre- (NP-40) to obtain soluble and insoluble fractions. left panel, immunoblot bleaching fluorescence ratio of Al546/Cy5. (f) IB3-1 cells were transfected analysis of GFP. B: core-glycosylated form (MW approximately 150 kDa, with either human HA-tagged-beclin 1 or empty vector. FRET analysis of TG2- revealed as 190kDa as GFP-tagged). right panel, densitometric analysis (mean Alexa 546 fluorescence after SUMO-1 Cy5 photobleaching. (g) IB3-1 cells ± s.d. of three independent experiments). (b) Immunoblot of E-cadherin as were transfected with either human Beclin 1 or scrambled oligonucleotides. control of membrane fractioning. (c-d) IB3-1 cells transfected with either Immunoblot analysis of beclin1. β-actin was used as loading control. (h) IB3-1 human HA-tagged-beclin 1 or empty vector. (c) TNF-α secretion. Each bar cells transfected with 50 nM human ATG5 siRNA or scrambled oligonucleotides represents the mean ± s.d. of three independent experiments. *, p<0.01 or cultured with cystamine in presence or absence of 3MA. FRET analysis of vs the empty vector, ANOVA. (d) Intracellular ROS content [CM-H2DCFDA TG2-Alexa 546 fluorescence after SUMO-1-Cy5 photobleaching. F/F0 indicates (DCF) detection by a Wallac 1420 multilabel counter]. Each bar represents post-bleaching/pre-bleaching fluorescence ratio of Al546/Cy5.

8 8

8 8 8

8 8 8 88 8 8 8 8

8 γ

8 8 8 8

8 8

8 88

Figure S7 Defective autophagy in human nasal mucosa and in CF mouse absence of 3-MA followed by rosiglitazone (d) FRET analysis of TG2-Alexa models. (a) Confocal microscopy micrographs of p62 in control and CF 546 fluorescence after SUMO-1 Cy5 photobleaching in lung tissues from human nasal polyp mucosa. DAPI (blue), nuclear counterstaining. Scale CftrF508del homozygous mice treated with cystamine or PBS in presence or bar, 10 m. Representative of five control and five CF patients (b) FRET absence of 3-MA (representative of seven CftrF508del mice). (e) Fluorescence- analysis of TG2-Alexa 546 fluorescence after SUMO-1 Cy5 photobleaching microscopy quantitation of LC3 dots in lung tissues from WT and Scnn1b-Tg in human nasal mucosae treated with cystamine or NAC in presence or mice (10 mice per group). Mean ±s.d. *, p<0.01 compared to WT mice. (f) absence of 3-MA. Representative of five cultured CF human nasal polyp Confocal microscopy micrographs of p62 in WT and Scnn1b-Tg mice. DAPI mucosa (patients #6-10, Supplementary Table 1 online). (c) FRET analysis (blue), nuclear counterstaining. Scale bar, 10 m. (g) Confocal microscopy of PPAR -Alexa 546 fluorescence after N-CoR-Cy5 photobleaching in micrographs of TG2 activity in WT and Scnn1b-Tg mice. DAPI (blue), in human nasal mucosae treated with cystamine or NAC in presence or nuclear counterstaining. Scale bar, 100 m.

Figure S8 Schematic representation of the main findings of the study. a: of F508del-CFTR in HDAC6+/p62+ aggresomes. (e) Defective autophagy Defective autophagy in CF airway epithelia. (a) Defetive CFTR and mutant generates a vicious cycle (black arrows) with increasing ROS and TG2 levels CFTR-induced ER stress both increase ROS levels and drive TG2 SUMOylation and (f) inflammation. b: Effects of cystamine treatment. (a) Cystamine inhibits and persistence of high TG2 protein levels. (b) ROS-driven TG2 SUMOylation TG2 cross-linking activity and also decreases TG2 SUMOylation, by reducing induces cross-linking, ubiquitination, and aggresome sequestration of PPAR ROS levels, favouring TG2 ubiquitination and proteasome degradation. (b) This and IK-B . TG2 also cross-links beclin 1 and favours its sequestration avoids cross-linking of beclin 1 leading to restoration of beclin 1 interactome in aggresomes. Cross-linked beclin 1 dislocates PI3K complex III away at the ER levels and allows autophagosome formation and autophagy upon from the Endoplasmic Reticulum (ER) leading to sequestration of beclin 1 starvation. (c) Restoration of beclin 1 and autophagy decreases p62 and interactome into aggresomes. (c) This leads to inhibition of autophagy. (d) aggresome formation and also leads to (d) restoration of PPAR and IK-B , Defective autophagy induces accumulation of damaged mitochondria and thus dampening inflammation. The inhibition of aggresome sequestration of p62 accumulation that leads to proteasome overload, favours the formation misfolded/modified proteins also allows F508delCFTR to traffic to the Golgi of protein aggregates, impairs CFTR trafficking leading to sequestration network and plasma membrane.

Figure S9 Densitometric analysis of blot experiments and quantitation of confocal microscopy micrographs.

66 6 6 6

6 6 6

6 6

Figure S10 Full scan of key blot experiments

Supplementary Tables

Supplementary Table 1: Clinical characteristics of Cystic Fibrosis patients.

Supplementary Table 2: Sequence of primers used for Quantitative RT-PCR and sequence of siRNA oligonucleotides

Supplementary Table 3: Primary antibodies

Patients # 1 2 3 4 5 6 7 8 9 10

Sex; F M F M F M M F M F

Age* 13, 2 14, 3 13,4 13, 1 13, 6 13,3 29 19 11 24

Age at diagnosis* 0, 6 0, 3 0, 8 2, 3 11, 0 1,5 7,2 2,0 0,4 7,0

Genotype F508del/ F508del/ F508del/ F508del/ F508del/ F508del/ F508del/ F508del/ F508del/ F508del/ F508del W1282X N1303K G542X F508del F508del G542X W1282X F508del W1282X

Pancreatic Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes insufficiency

Chronic Yes No No No Yes Yes No Yes Yes No respiratory infection (PA)

Mean FEV1, % 69 78 73 80 69 81 72 64 72 75 predicted

# patient's number; *, (years,months)

Sequence of primers used for Quantitative RT-PCR

GENE NAME FOWARD PRIMER SEQUENCE (5'-3') REVERSE PRIMER SEQUENCE (5'-3') β-actin ATTGCCGACAGGATGCAGAA ACATCTGCTGGAAGGTGGACAG ULK1 GTCCAGCTTTGACAGTCAGT TCCCAGGACTCAGGATTCCA ULK2 TGTTTCTAGGGTGTAGCTGG CACCATCTACCCCAAGACCT ATG5 AGTATCAGACACGATCATGG TGCAAAGGCCTGACACTGGT ATG6\BECLIN1 TGTCACCATCCAGGAACTCA CTGTTGGCACTTTCTGTGGA ATG7 CACTGTGAGTCGTCCAGGAC CGCTCATGTCCCAGATCTCA ATG8 CGTGGAGTCCGCGAAGATTC AGGCATGAGGACAATGCACA ATG14L ATGAGCGTCTGGCAAATCTT CCCATCGTCCTGAGAGGTAA Bcl2 GCTGCTGGAGGCTGGGGAGA CCCCCACAGGAACCCTCCTC hVsps34 AAGCAGTGCCTGTAGGAGGA TGTCGATGAGCTTTGGTGAG hVsps15 TGGGACTGGCGGTCTGCACT CGGGACGGGAAAACGCGTCA Ambra-1 AACCCTCCACTGCGAGTTGA TCTACCTGTTCCGTGGTTCTCC

Sequence of siRNA oligonucleotides

GENE NAME OLIGO-SENSE(5'-3') OLIGO-ANTISENSE (5'-3')

Control UACCGUCUCCACUUGAUCGdTdT CGAUCAAGUGGAGACGGUAdTdT TG2 GCGUCGUGACCAACUACAAdTdT UUGUAGUUGGUCACGACGCdGdG PIASγ oligo#2 CAAGACAGGUGGAGUUGAUUU pAUCAACUCCACCUGUCUUGUU PIASγ oligo#4 AAGCUGCCGUUCUUUAAUAUU pUAUUAAAGAACGGCAGCUUUU BECLIN-1 AAGAUUGAAGACACAGGAGGC GCCUCCUGUGUCUUCAAUCUU p62 UACAAAUUUAACAGGAUGGdTdT CCAUCCUGUUAAAUUUGUAdTdT CFTR oligo#1 CGUGUGUCUGUAAACUGAUGGCUAA UUAGCCAUCAGUUUACAGACACAGC CFTR oligo#2 CCCUUCUGUUGAUUCUGCUGACAAU AUUGUCAGCAGAAUCAACAGAAGGG CFTR oligo#3 GGCAUAGGCUUAUGCCUUCUCUUUA UAAAGAGAAGGCAUAAGCCUAUGCC ATG5 oligo#1 GGUUUGGACGAAUUCCAACUUGUUU AAACAAGUUGGAAUUCGUCCAAACC ATG5 oligo#2 GAUCACAAGCAACUCUGGAUGGGAU AUCCCAUCCAGAGUUGCUUGUGAUC ATG5 oligo#3 CAAAGAAGUUUGUCCUUCUGCUAUU AAUAGCAGAAGGACAAACUUCUUUG

ANTIBODY COMPANY APPLICATION phospho-p42/p44 MAP kinases CELL SIGNALING IB E-cadherin CELL SIGNALING IB β-actin (13E5) CELL SIGNALING IB PPARγ (E8) SANTA CRUZ BIOTECNOLOGY IB PIASγ(H75) SANTA CRUZ BIOTECNOLOGY IB Ubiquitin(FL76) SANTA CRUZ BIOTECNOLOGY IB, IF TG2 (H237) SANTA CRUZ BIOTECNOLOGY IB TG2 (CUB7402) NEOMARKER IB GFP(FL) SANTA CRUZ BIOTECNOLOGY IB, IF hVps15\p150 (Nterm) SANTA CRUZ BIOTECNOLOGY IB, IF Bcl-2(C2) SANTA CRUZ BIOTECNOLOGY IB, IF HDAC-6 SANTA CRUZ BIOTECNOLOGY IF SUMO-1 SANTA CRUZ BIOTECNOLOGY IF COX-2 SANTA CRUZ BIOTECNOLOGY IF BECLIN-1 ABCAM IB,IF LC3 ABCAM IB,IF phospho-eIF2α ABCAM IB CFTR ABCAM IB,IF AMBRA-1 ABCAM IB p62 BD BIOSCIENCE IB,IF HA BD BIOSCIENCE IB αβ-TUBULIN SIGMA IB hVps34 SIGMA IB,IF Atg14L SIGMA IB,IF LAMP1 SIGMA IF Nε(γ-L-glutamyl)-L-lysine isopeptide (81DIC2) COVOLAB IB CD68 ACRIS IF

FRET Alexa 546anti-TG2 / Cy5-anti anti-SUMO-1 (Molecular Probes, INVITROGEN) FRET Alexa 546anti-beclin 1 / Cy5-anti anti-HDAC6 (Molecular Probes, INVITROGEN)

Alexa 546anti-PPARγ/ Cy5-anti anti-N-CoR (Molecular Probes, INVITROGEN)/SANTA CRUZ FRET FRET Alexa 546anti-Bcl-2/ Cy5-anti anti-Beclin (Molecular Probes, INVITROGEN) FRET Alexa 546anti-Beclin 1 / Cy5-anti Vps34 (Molecular Probes, INVITROGEN) FRET Alexa 546anti-Beclin 1 / Cy5-anti Vps34 (Molecular Probes, INVITROGEN) FRET Alexa 546anti-Beclin 1 / Cy5-anti Vps15 (Molecular Probes, INVITROGEN) FRET Alexa 546anti-Beclin 1 / Cy5-anti Ambra-1 (Molecular Probes, INVITROGEN)

Alexa FRET 546anti-Beclin 1 / Cy5-anti ATG14L (Molecular Probes, INVITROGEN) FRET Alexa 546anti-CFTR / Cy5-anti HDAC-6 (Molecular Probes, INVITROGEN) FRET Alexa 546-anti GFP / Cy5-anti HDAC-6 (Molecular Probes, INVITROGEN)

Lysosomal accumulation of gliadin p31−43 peptide induces oxidative stress and tissue transglutaminase-mediated PPARγ downregulation in intestinal epithelial cells and coeliac mucosa

Luigi Maiuri, Alessandro Luciani, Valeria Rachela Villella, et al.

Gut 2010 59: 311-319 originally published online December 1, 2009 doi: 10.1136/gut.2009.183608

Updated information and services can be found at: http://gut.bmj.com/content/59/3/311.full.html

These include: References This article cites 52 articles, 18 of which can be accessed free at: http://gut.bmj.com/content/59/3/311.full.html#ref-list-1

Article cited in: http://gut.bmj.com/content/59/3/311.full.html#related-urls Email alerting Receive free email alerts when new articles cite this article. Sign up in the service box at the top right corner of the online article.

Notes

To order reprints of this article go to: http://gut.bmj.com/cgi/reprintform

To subscribe to Gut go to: http://gut.bmj.com/subscriptions Downloaded from gut.bmj.com on June 25, 2010 - Published by group.bmj.com

Coeliac disease Lysosomal accumulation of gliadin p31e43 peptide induces oxidative stress and tissue transglutaminase-mediated PPARg downregulation in intestinal epithelial cells and coeliac mucosa Luigi Maiuri,1 Alessandro Luciani,1,2 Valeria Rachela Villella,3 Angela Vasaturo,3 Ida Giardino,4 Massimo Pettoello-Mantovani,1 Stefano Guido,2,3 Olivier N Cexus,5 Nick Peake,5 Marco Londei,6 Sonia Quaratino5

See Commentary, p 286 ABSTRACT been found to be ‘toxic’ for predisposed individuals < Supplementary methods and Background An unresolved question in coeliac disease is and it has also been suggested that different portion(s) figures are published online only to understand how some toxic gliadin peptides, in of gliadin, ostensibly not those recognised by Tcells, at http://gut.bmj.com/content/ particular p31e43, can initiate an innate response and modulates this innate activation of coeliac small vol59/issue3 lead to tissue transglutaminase (TG2) upregulation in intestine.78Among these, p31e43 induces the 1Institute of Pediatrics, coeliac intestine and gliadin sensitive epithelial cell lines. expression of early markers of epithelial activation University of Foggia, Foggia, Italy Aim We addressed whether the epithelial uptake of and leads to enterocyte apoptosis4 both in coeliac 2Dynamic Imaging Microscopy, p31e43 induces an intracellular pro-oxidative intestine and in sensitive intestinal epithelial cell CEINGE, Naples, Italy 4 3Department of Chemical envoronment favouring TG2 activation and leading to the lines. Engineering, University Federico innate immune response. The mechanisms by which p31e43 initiates an II of Naples, Naples, Italy Methods The time course of intracellular delivery to innate immune response in coeliac duodenum are, 4 Department of Biomedical lysosomes of p31e43, pa-2 or pa-9 gliadin peptides was however, still elusive. It has been reported that Science, University of Foggia, e Foggia, Italy analysed in T84 and Caco-2 epithelial cells. The effects of p31 43 can delay epidermal growth factor receptor 5Cancer Research UK Oncology peptide challenge on oxidative stress, TG2 and (EGFR) inactivation through interference with the Unit, University of Southampton, peroxisome proliferator-activated receptor (PPAR)g endocytic pathway,9 this suggesting that gliadin Southampton, UK e 6 ubiquitination and p42/44 mitogen activated protein fragments could amplify the effects of trace Novartis Pharma AG (MAP) kinase or tyrosine phosphorylation were e Translational Medicine, Basel, amounts of EGF. Moreover, p31 43 induces early Switzerland investigated in cell lines and cultured coeliac disease tissue transglutaminase (TG2) upregulation biopsies with/without anti-oxidant treatment or TG2 gene compared to immunodominant gliadin peptides.10 Correspondence to silencing by immunoprecipitation, western blot, confocal This indicates that TG2 is not only central to the Professor Luigi Maiuri, Institute microscopy and Fluorenscence Transfer Resonance adaptive response to gliadin as the main deami- of Pediatrics, University of Foggia, viale Pinto 1, Foggia Energy (FRET) analysis. dating enzyme of the immunodominant epitopes, 71100, Italy; [email protected] Results After 24 h of challenge p31e43, but not pa-2 or but is also a key player in the initiation of inflam- pa-9, is still retained within LAMP1-positive perinuclear mation in coeliac disease.10 However, it is still Revised 13 July 2009 vesicles and leads to increased levels of reactive oxygen unknown how p31e43 induces early TG2 activa- Accepted 1 September 2009 species (ROS) that inhibit TG2 ubiquitination and lead to tion in both coeliac intestine and ‘gliadin-sensitive’ Published Online First 910 1 December 2009 increases of TG2 protein levels and activation. TG2 intestinal epithelial cells. induces cross-linking, ubiquitination and proteasome TG2 expression is regulated by retinoids, steroid degradation of PPARg. Treatment with the antioxidant hormones, peptide growth factors and cytokines, EUK-134 as well as TG2 gene silencing restored PPARg which also lead to a time-dependent decrease in levels and reversed all monitored signs of innate TG2 ubiquitination,11 indicating that post-trans- activation, as indicated by the dramatic reduction of lational control mechanisms also play a role in TG2 tyrosine and p42/p44 phosphorylation. regulation. We have demonstrated that in cystic Conclusion p31e43 accumulation in lysosomes leads to fibrosis (CF) airway epithelia,12 sustained TG2 epithelial activation via the ROSeTG2 axis. TG2 works as activation is driven by the increased levels of reac- a rheostat of ubiquitination and proteasome degradation tive oxygen species (ROS) caused by the CFTR and drives inflammation via PPARg downregulation. genetic defect.12 We therefore investigated whether p31e43 was able to drive an intracellular pro- oxidative environment and demonstrate that p31e43 leads to an increase of ROS levels in coeliac INTRODUCTION duodenum as well as T84 and Caco-2 intestinal Coeliac disease is a relatively common pathology epithelial cells. that affects 1% of the population, and is precipi- To understand how p31e43 challenge induces e tated by the ingestion of gluten proteins.1 3 In oxidative stress, we analysed the delivery of this predisposed individuals, gliadin acts as a stimulator peptide to the intracellular degradation systems. It has of innate responses, in turn setting the type and been reported that the epithelial translocation of the intensity of the adaptive immune response4 leading gliadin immunostimulatory peptide 33-mer occurs by to a powerful activation of lamina propria gliadin- transcytosis after partial degradation through a rab5 e specific Tcells.5 7 Several fragments of gliadin have endocytic compartment.13 Increased transepithelial

Gut 2010;59:311e319. doi:10.1136/gut.2009.183608 311 Downloaded from gut.bmj.com on June 25, 2010 - Published by group.bmj.com

Coeliac disease translocation of the 33-mer was demonstrated in active coeliac uals, and the study was performed according to the Ethics disease compared to healthy controls. Moreover, incomplete degra- Committee of Regione Campania Health Authority. One spec- dation of the 33-mer and protected transport of the peptide 31e49 imen from each patient was used for diagnosis; the other samples occurs in patients with untreated coeliac disease, thus favouring were cultured in vitro as described.4 their respective immunostimulatory and toxic effects.14 Here we show that p31e43 is retained within the lysosomes as compared to In vitro organ culture of biopsy specimens from patients with the more rapid disappearance of the immunodominant peptides coeliac disease and from controls pa-2 or pa-9 from the intracellular compartments. Blocking p31e43 Duodenal biopsy specimens were cultured as previously endocytosis prevents the increase of ROS. reported4 for 3 or 24 h with medium alone, p31e43, pa-9, pa-2, We have previously reported that in CF airways sustained pTPO (20 mg/ml) in the presence or absence of cystamine TG2 activation drives inflammation by inducing cross-linking, (400 mmol/l). Duodenal biopsy specimens were also incubated for ubiquitination and proteasome degradation, of the anti-inflam- 6 h with the PPARg agonist rosiglitazone (10 mmol/l; Alexis matory peroxisome proliferator-activated receptor (PPAR)g,12 Biochemicals) and then re-challenged for 3 h with p31e43. a negative regulator of inflammatory gene expression.15 16 Blocking TG2 through specific gene silencing12 or the specific RNA interference TG2 inhibitors restores a physiological control of inflammation The cells were transfected with 50 nmol/l human TG2 and in CF airways.12 scrambled small interfering RNAs (siRNAs) duplex using We therefore addressed whether the p31e43-induced innate Hiperfect Transfection Reagent (Qiagen, Milan, Italy) at 378C for response was mediated by the ROSeTG2 axis via PPARg 72 h as previously described.12 downregulation. Adenoviral vector Human manganese superoxide dismutase (MnSOD) cDNA was METHODS cloned into the shuttle vector pAd5CMVK-NpA.18 MnSOD Peptide preparation adenovirus was a gift from Michael Brownlee (Albert Einstein Biotinylated gliadin peptides p31e43, pa-9 (57e68) or pa-2 College of Medicine, New York). T84 cell lines were infected with (62e75), control human thyroid peroxidase pTPO (535e551) MnSOD or the control adenovirus for 2 h, as previously and pTPO (536e547), peptide LGQQQPFAAVQPY (referred to described.18 below as pZ), and scrambled peptides with amino acid compo- e sition similar to p31 43 (QQGQPFPQPQLQY (referred to below Immunolocalisations as pX), GLQQFQPPPPQQY (referred to below as pY)) were Tissue sections were individually incubated with the antibodies synthesised by Primm (Milan, Italy) and the purity was deter- anti-phospho-tyrosine (PY99 mAb, 1:80, mouse IgG2b; Santa mined by reverse-phase HPLC. Cruz Biotechnology, Santa Cruz, California, USA), anti-PPARg (clone E8, sc-7273, 1:100, mouse IgG1; clone H100, sc-7196, Cell lines 1:100, rabbit polyclonal IgG; Santa Cruz Biotechnology) and Human colon adenocarcinoma T84 or human colorectal carci- control antibodies.4 12 Cell lines were incubated with the anti- noma Caco-2 cell lines were cultured as recommended by bodies against phospho-tyrosine, Early Endosome Antigen-1 American Type Culture Collection (ATCC). (EEA-1; 1:100, rabbit polyclonal IgG; Abcam, Cambridge, UK), Rab7 (1:200, rabbit polyclonal IgG; Abcam), ARF-1 (1:100, rabbit Cell cultures polyclonal IgG; Abcam), LAMP-1 (1:100, rabbit polyclonal IgG; The cells were challenged with p31e43 for 24 h in the presence Abcam). Two-colour and indirect immunofluorescence were used or absence of the ROS scavenger EUK-134 (50 mg/ml; Alexis for the detection of the tested markers by using an LSM510 Zeiss Biochemicals, Florence, Italy), glutathione (GSH) (10 mmol/l; confocal laser scanning unit (Zeiss, Jena, Germany).12 The Sigma, Milan, Italy)17 or the TG inhibitor cystamine (400 mmol/l; correlation coefficient analysis was used as value of co-local- Sigma). To study internalisation of peptides, cells were chal- isation between peptides and EEA-1 or LAMP-1 as described.19 lenged for 15, 30, 60 or 90 min, 3 or 24 h at 378C with the biotinylated peptides (20 mg/ml). To inhibit endocytosis, the cells In situ detection of TG2 enzymatic activity were pre-treated with the cholesterol-binding agent methyl-b- The detection of in situ TG2 activity was performed by detec- cyclo-dextrin (M-b-CD, 10 mmol/l; Sigma) or filipin (5 mg/ml; tion of biotinylated mono-dansyl-cadaverine (MDC) incorpora- Sigma) and then challenged with biotinylated peptides. Cell lines tion by Alexa-546 streptavidin as previously described.4 12 were also incubated for 6 h with the PPARg agonist rosiglitazone (10 mmol/l; Alexis Biochemicals) and then challenged with FRET microscopy p31e43. The PPARg antagonist GW9662 (1 mmol/l; Alexis For acceptor photobleaching 4-mm frozen tissue sections of biopsy Biochemicals) was added for 24 h to p31e43 in the presence or samples were fixed with buffered 2% paraformaldehyde and absence of TG2 siRNA oligos. The proteosome inhibitor MG132 permeabilised with 0.5% Triton X-100. Upon fixation, tissues were (50 mmol/l; Calbiochem, Milan, Italy) was added 6 h before immunostained with Alexa 546/anti-PPARg (Santa Cruz p31e43 in the presence or absence of TG2 siRNA oligos or Biotechnology, Santa Cruz, California, USA) and Cy5/anti-Ne(g-L- EUK134. glutamyl)-L-lysine isopeptide (Covolab, Cambridge, UK). Alexa 546 fluorescence was detected before and after Cy5 photobleaching. Patients Duodenal multiple endoscopic biopsies were performed for Immunoblot and immunoprecipitation diagnostic purposes in seven treated patients with coeliac disease Blots were incubated with anti-phospho-tyrosine (PY99 mAb, (mean age, 25 years; range, 16e41 years), and five non-coeliac 1:500), anti-phospho-p42/p44 MAP kinases (Cell Signaling controls affected by oesophagitis (22.4 years; range, Technology, Danvers, Massachusetts, USA), PPARg (clone E8 sc- 18e29 years). Informed consent was obtained from all individ- 7273; Santa Cruz Biotechnology, Santa Cruz, California, USA),

312 Gut 2010;59:311e319. doi:10.1136/gut.2009.183608 Downloaded from gut.bmj.com on June 25, 2010 - Published by group.bmj.com

Coeliac disease

TG2 (clone CUB7402, 1:100; NeoMarkers, Fremont, California, and pY peptides, behaved like pa-9 and pa-2 in both T84 USA), Ne(g-L-glutamyl)-L-lysine isopeptide (clone 81DIC2; (supplementary figure 1) and Caco-2 cell lines (data not shown). Covalab), ubiquitin (1:100 clone FL-76, rabbit polyclonal IgG; e Santa Cruz Biotechnology), as previously described.12 p31 43 induces increase of ROS in T84 epithelial cells Immunoprecipitations with anti-TG2 CUB 7402 mAb, anti- T84 cells were pulsed with 10 mmol/l CM-H2DCFDA in the phospho-tyrosine and anti-PPARg were carried out as previously presence of the tested peptides and the levels of ROS were described.12 monitored after 90 min, 3 h and 24 h of challenge. The intracellular transport of gliadin peptides through early and late endo- ROS detection cytic endosomes did not induce ROS generation (data not Cell lines were pulsed with 10 mmol/l 5-(and-6)-chloromethyl- shown). Moreover, after 90 min of challenge neither p31e43 nor 29,79-dichlorodihydrofluorescein diacetate acetyl ester (CM- pa-9, both detected in LAMP-1 positive vesicles (figure 1C), H2DCFDA) (Molecular Probes, Invitrogen, Milan, Italy) and induced significant increase of ROS (figure 2A). At the later time analysed as previously described.12 points (3e24 h of challenge), p31e43, but not pa-9 (figure 2A), pa-2, pTPO (535e551) (data not shown) or peptides pTPO Statistical analysis (536e547), pX, pY and pZ (supplementary figure 1) induced Cells challenged with peptides were compared with those a time-dependent significant increase of ROS levels, suggesting cultured in medium alone, and with those cultured in the that the prolonged persistence of p31e43 in LAMP-1 positive presence of the tested inhibitors. All experiments were vesicles (figure 1E,F) generates a pro-oxidative environment. No performed at least in triplicate. Data distribution was analysed, increase of ROS was observed after challenge with p31e43 upon and statistical differences were evaluated using ANOVA M-b-CD or filipin treatment (figure 2B). TukeyeKramer test and SPSS 12 software. A p value of <0.05 The generation of a pro-oxidative environment induces acti- was considered significant. vation of different stress sensitive signalling pathways.9 12 21 Detailed methods are described in the online supplementary Since the challenge with p31e43 induces phosphorylation of the information section. extracellular signal-regulated kinases 1/2 (Erk1/2, p42/44 mitogen-activated protein kinases, MAPK),9 we investigated RESULTS whether p31e43-induced ROS-mediated p42/44 MAPK phos- p31 e43 is retained within the lysosomes in T84 epithelial cells phorylation in T84 cells. We demonstrated that p42e44 MAPK e T84 cells were incubated with biotinylated p31e43, p-a2, p-a9 phosphorylation induced by the challenge with p31 43 was e (20 mg/ml) for 15, 30, 60, 90 min, 3 h and 24 h at 378C. The prevented by the incubation with the catalase superoxide 12 fi peptides were rapidly detected (upon 15 min) in intracellular dismutase (SOD) mimetic EUK-134 ( gure 2C), as well as by 17 18 fi vesicles (figure 1A). Prolonged incubation (after 90 min to 24 h) GSH or by MnSOD over-expression (supplementary gure 2). increased the amount of p31e43 (figure 1A), but not of p-a9 This further indicates that the generation of a pro-oxidative e (figure 1A) or p-a2 (data not shown), which resided in perinuclear environment is critical to p31 43 biological activity in T84 cells. vesicles. Soon after 15 min of challenge, peptides co-localised p31e43 induces ROS-dependent inhibition of TG2 ubiquitination with EEA-1 (figure 1B), a marker of early endosomes,13 whereas To investigate whether the increased levels of ROS may account after 60 min they co-localised with Rab7-positive late endosomes for the p31e43-induced increase of TG2 levels,12 we incubated (data not shown).20 These data are in agreement with those p31e43-challenged T84 cells with EUK-134. We demonstrated described for the gliadin peptide 33-mer.13 Co-staining with anti- that EUK-134 was effective in controlling the increase of TG2 body against Arf1 (data not shown) did not reveal translocation protein levels (figure 3A) and TG2 activity (data not shown) of the peptides to the Golgi complex. After 90 min the internalised induced by p31e43. The same effects as EUK-134 were observed peptides co-localised with Lysosomal Associated Membrane after incubation with GSH or upon MnSOD over-expression Protein-1 (LAMP-1) (figure 1C), a marker of lysosomes.20 The (supplementary figure 2). No increase of TG2 was observed after peptide/EEA-1 or peptide/LAMP-1 correlation coefficients of the challenge with p31e43 upon M-b-CD or filipin treatment (data spatial intensity distributions yielded a fairly high correlation not shown). This indicates that p31e43 internalisation is critical coefficient (figure 1D), indicating that gliadin peptides are for the induction of ROS-mediated TG2 activation. internalised and transported through early and late endocytic Since post-translational modifications of TG2, such as ubiq- endosomes to lysosomes. No internalisation of peptides or co- uitination, play a role in regulating the levels of TG2 protein11 localisation with LAMP-1 were observed when T84 cells were we investigated whether TG2 ubiquitination was influenced by pre-incubated with M-b-CD, which inhibits clathrin-mediated ROS generation induced by p31e43. We incubated T84 cells with endocytosis,20 or filipin, a specific inhibitor of lipid raft- or p31e43 in the presence or absence of EUK-134 upon inhibition of caveolae-dependent endocytosis20 (data not shown). This proteasome function by MG132,12 then immunoprecipitated suggests that endocytosis is essential for the delivery of gliadin TG2 protein and detected ubiquitin co-reactivity. The ubiquitin peptides to the lysosomes for degradation. immunoreactivity on TG2 immunoprecipitates was enhanced Increasing appearance of the internalised p31e43 in LAMP-1- upon treatment with EUK-134 (figure 3B). No effects of p31e43 positive vesicles, mainly limited to perinucelar localisation, was on TG2 ubiquitination were observed when T84 cells were observed after 3e24 h of challenge (figure 1E,F). At these late pre-treated with M-b-CD or filipin (data not shown). time points p-a9(figure 1E,F) or pa-2 (data not shown) were only faintly detected in LAMP-1-positive structures. The coefficients p31e43-induced TG2 drives PPARg cross-linking and of p-a2orp-a9 co-localisation with LAMP-1 were significantly proteasome degradation in T84 cells lower compared to p31e43 (figure 1G), thus indicating that We investigated whether the p31e43-induced TG2 activation p31e43, but not p-a9orp-a2 is retained within the lysosomes. was effective in inducing PPARg cross-linking and proteasome Caco-2 cells showed the same behaviour as T84 after peptide degradation as we have reported in CF airways.12 challenge (data not shown). The control pTPO (535e551) (data We immunoprecipitated PPARg species from T84 cell lysates not shown), pTPO (536e547) and pZ, as well as scrambled pX after challenge with p31e43 and detected the immunoreactivity

Gut 2010;59:311e319. doi:10.1136/gut.2009.183608 313 Downloaded from gut.bmj.com on June 25, 2010 - Published by group.bmj.com

Coeliac disease

Figure 1 p31e43, but not pa-9, accumulates within the lysosomes in T84 cells. (A) Incubation of T84 cells with biotinylated p31e43 and pa-9 from 15 min to 24 h at 378C and detection by confocal microscopy upon incubation with Alexa 488-conjugated streptavidin. After 15 min both p31e43 and p-a9 were detected in intracellular peripheral vesicles. Prolonged incubation (90 min to 24 h) increased the amount of p31e43, but not of p-a9, in perinuclear vesicles. (B) Co-localisation of p31e43 and p-a9 (green) with EEA-1 (red) in T84 cells upon 15 min of challenge. Both peptides were detected in EEA1-positive vesicles (yellow). (C) Co-localisation of p31e43 and p-a9 (green) with LAMP-1 (red) in T84 cells after 90 min of challenge. Both peptides were detected in LAMP1-positive vesicles (yellow). (D) Quantitative measurement of peptides/EEA-1 or peptides/LAMP-1 co-localisations shown in B and C, respectively. Each bar represents the mean plus SEM of three independent experiments. (EeF) Co-localisation of p31e43 and p-a9 (green) with LAMP-1 (red) in T84 cells after 3 h and 24 of challenge. p31e43, but not pa-9, was detected in LAMP-1 positive vesicles (yellow). (G) Quantitative measurement of peptides/LAMP-1 co-localisations shown in E and F, respectively. Each bar represents the mean plus SEM of three

314 Gut 2010;59:311e319. doi:10.1136/gut.2009.183608 Downloaded from gut.bmj.com on June 25, 2010 - Published by group.bmj.com

Coeliac disease

Figure 2 p31e43 induces increase of ROS levels in T84 cells. (A) Increase of intracellular ROS in p31e43 stimulated T84 cells. Each bar represents the mean plus SEM of three independent experiments. *p<0.01 versus samples cultured with medium or with p-a9. (B) Inhibition of p31e43 endocytosis by methyl-b-cyclo-dextrin or ﬁlipin inhibited p31e43-induced increase of ROS. Each bar represents the mean plus SEM of three independent experiments. *p<0.017 versus samples cultured with p31e43. (C) Immunoblot analysis showed a decrease of p31e43-induced p42/p44 phosphorylation upon treatment with the ROS scavenger EUK-134. Representative result of three different experiments. DCF, 5-(and-6)- chloromethyl-2‘,7’ -dichlorodihydroﬂuorescein diacetate acetyl ester; ROS, reactive oxygen species.

with an isopeptide cross-link specifically catalysed by TG2.22 (figure 4B). Negligible amounts of ubiquitinated PPARg were We observed isopeptide immunoreactivity in PPARg immuno- observed in p31e43-stimulated T84 cells upon TG2 gene precipitates (figure 4A). No PPARg cross-linking was observed in silencing (figure 4B) as well as after TG2 inhibition by cystamine T84 cells challenged with pa-9 or pa-2 (data not shown). TG2 or after treatment with EUK-134 (data not shown). Moreover, siRNA12 was highly effective in preventing PPARg cross-linking a reduction of the 55-kDa PPARg form was observed in T84 cells upon p31e43 exposure (figure 4A). Moreover, pre-treatment upon p31e43 challenge (figure 4C) and TG2 siRNA was highly with the TG2 inhibitor cystamine,23 as well as with the EUK- effective in preventing p31e43-induced PPARg downregulation 134, also prevented TG2-mediated cross-linking of PPARg (data (figure 4C and supplementary figure 3). The incubation with the not shown). These results indicate that p31e43 induces post- TG2 inhibitor cystamine (400 mmol/l), showed the same effects translational modifications of PPARg via the ROSeTG2 axis. as TG2 knock-down (figure 4C). No PPARg cross-linking or We also investigated whether the cross-linking of PPARg proteasome degradation was observed in T84 cells upon inhibi- might favour ubiquitination and PPARg proteasome degradation tion of p31e43 endocytosis by M-b-CD or filipin (data not with reduction of the 55 kDa PPARg form. Indeed, a striking shown). These results indicate that the internalisation and increase of ubiquinated PPARg was observed in p31e43-stimu- lysosomal delivery of p31e43 induces PPARg downregulation via lated T84 cells treated with the proteasome inhibitor MG132 the ROSeTG2 axis. independent experiments. (AeC,E,F), representative results of three independent experiments. Confocal microscopy, DAPI (blue) nuclear counterstaining. Scale bar, 10 mm. DAPI, 49-diamidine-29-phenylindol dihydrochloride; EEA-1, Early Endosome Antigen 1; LAMP-1, Lysosomial Associated Membrane Protein 1.

Gut 2010;59:311e319. doi:10.1136/gut.2009.183608 315 Downloaded from gut.bmj.com on June 25, 2010 - Published by group.bmj.com

Coeliac disease

regulation was observed in non-coeliac control biopsies (figure 5C), in agreement with our previous data showing that p31e43 fails to induce TG2 upregulation in non-coeliac duodenum.10 Moreover, upon challenge with p31e43 coeliac biopsies showed intense immunoreactivity to the anti-phospho-tyrosine antibody PY99 (figure 5D). Pretreatment with rosiglitazone completely reversed the p31e43-induced phospho-tyrosine immunoreactivity (figure 5D).

DISCUSSION An innate response triggered by several ‘toxic’ gliadin fragments4 78is involved in both modulating mucosal e damage24 26 as well as in setting the type and intensity of the Figure 3 ROS-mediated inhibition of TG2 ubiquitination in T84 cells adaptive immune response in coeliac disease.4 Furthermore, a series upon p31e43 challenge. (A) Immunoblot analysis of TG2 protein. ROS of studies has indicated that several gliadin fragments activate, in a scavenger EUK-134 inhibited p31e43-induced increase of TG2 protein in non-disease-specific manner, both mouse and human DC cells,27 28 T84 cells. (B) Ubiquitin immunoreactivity in TG2 immunoprecipitates. a specific type of DC detected in the intestines of patients with Incubation with EUK-134 increased TG2 ubiquitination upon MG132 29 coeliac disease and some epithelial cells lines derived from treatment in T84 cells. Immunoprecipitation (IP): anti-TG2 antibody; 30 31 10 immunoblot (IB): anti-ubiquitin antibody Representative results of three different species. These results have raised a series of questions and the most pressing is to understand why only CD independent experiments. ROS, reactive oxygen species; TG2, tissue 4 transglutaminase. tissues react to this non-T-cell antigenic portion of gliadin. Here we demonstrate that ‘gliadin sensitive’ epithelial cells upregulate intracellular ROS upon challenge with p31e43 but not with immunodominant gliadin peptides. These results indi- e e The ROS TG2 PPARg axis is a master regulator of the cate that p31e43 induces cellular stress. Among the factors able epithelial activation to gliadin peptides to trigger cellular stress is the accumulation of improperly e 32e35 To demonstrate that the biological activity of p31 43 was handled substances within the intracellular compartments. mediated by the downregulation of PPARg, we pre-incubated A string of studies has suggested that some a-gliadin peptides, in 12 T84 cells with the PPARg agonist rosiglitazone for 6 h and then particular p31e43(9)13 as well as 33-mer,14 possess the ability to e 13 challenged with p31 43. As a matter of fact, PPARg ligation by penetrate cells. It has been demonstrated that gliadin peptides agonists favours PPARg interaction with the nuclear receptor co- may be internalised by endocytic uptake and activate some signal repressor (N-CoR) histone deacetylase 3 (HDAC3) complex and transduction pathways.9 Here we demonstrate that p31e43 is thereby blocks its ubiquitination, thus maintaining a repressor delivered to the lysosomes but is retained as late as 24 h after 23 e condition. We monitored the effects of p31 43 on epithelial challenge in LAMP1-positive vesicles mainly located in the peri- tyrosine phosphorylation, as revealed by PY-99 antibody, a well- nuclear region. The engulfment of lysosomes with p31e43 may established marker of epithelial activation in T84 cells and in induce cellular stress with the generation of a pro-oxidative 4 human coeliac disease intestinal mucosa. We found that the pre- environment. The relationship between lysosomal garbage and e incubation with rosiglitazone antagonised p31 43-induced the perturbation of cellular homeostasis has been described in fi tyrosine phosphorylation in T84 cells ( gure 4D). a number of human pathologies such as lysosomal storage e 33 36 37 32 34 Moreover we incubated p31 43-challenged T84 cells with the diseases, neurodegenerative diseases and even in 12 PPARg antagonist GW9662 upon TG2 siRNA. The incubation ageing.35 However, in T84 or Caco-2 cells as well as in coeliac with GW9662 antagonised the downregulatory effect of the TG2 intestine, the lysosomal machinery of peptide degradation is not e siRNA on tyrosine phosphorylation induced by p31 43, as generally perturbed since other peptides, such as pTPO and even fi detected by immunoprecipitation studies ( gure 4E) and confocal the gliadin immunodominant pa-2 or pa-9 peptides, are no fi e microscopy ( gure 4F). These data indicate that p31 43 may longer detectable within intracellular organelles after 90 min of e e induce epithelial activation via the ROS TG2 PPARg axis. challenge. Whatever the mechanism involved in such a puzzling interaction between some gliadin peptides and ‘sensitive’ p31e43 induces TG2-mediated PPARg cross-linking and reduced epithelia, these results further indicate that a complex alteration protein expression in coeliac duodenum of the cross-talk between the intestine and its local environment To investigate whether TG2-mediated PPARg cross-linking and is crucial for the development of coeliac disease. downregulation occurs in human coeliac intestine, we chal- Our results demonstrate that p31e43 induces TG2 activation lenged coeliac biopsies with p31e43, pa-2 or pa-9 in presence or via ROS generation. The increased levels of ROS reduce TG2 absence of cystamine, a well known TG2 inhibitor already used degradation by the ubiquitineproteasome system, thus leading in vivo in a mouse model of Huntington’s disease to control to increased TG2 protein levels. TG2 is a multifunctional enzyme TG2-related manifestations.23 FRET analysis showed that with a vast array of biological functions.38 Increased tissue levels PPARg interacted with the isopeptide cross-link also in human of TG2 have been described in a number of human diseases, such coeliac disease biopsies (figure 5A). Confocal microscopy showed as neurodegenerative diseases,39 40 CF12 and even in cancers.41 that PPARg protein was reduced in coeliac biopsies after chal- The upregulation of TG2 has recently been associated with an lenge with p31e43 (figure 5B) but not with pa-2 or pa-9 (data increased metastatic activity41 or drug resistence.41 One of the not shown). Moreover treatment of cultured biopsies with TG2 effects is the cross-linking, with consequent functional cystamine was highly effective in preventing p31e43-induced sequestration and proteasome degradation of several intracellular PPARg downregulation (figure 5B). No PPARgeisopeptide proteins such as a-synuclein in Parkinson’s disease,39 huntingtin interaction (data not shown) or p31e43 induced PPARg down- in Huntington’s disease.42 We have previously reported that in

316 Gut 2010;59:311e319. doi:10.1136/gut.2009.183608 Downloaded from gut.bmj.com on June 25, 2010 - Published by group.bmj.com

Coeliac disease

Figure 4 p31e43 challenge induces TG2-mediated PPARg cross-linking and proteasome degradation in T84 cells. (A) Immunoprecipitates of PPARg species from whole-cell extracts of T84 cells after challenge with p31e43 were immunoreactive for the anti-isopeptide cross-link antibody. High MW bands ranging from 72 to 130 kDa are evident after p31e43 challenge and signiﬁcantly reduced upon TG2 siRNA treatment. Immunoprecipitation (IP): anti-PPARg antibody; immunoblot (IB): anti-isopeptide antibody. (B) Effect of the MG132 treatment on PPARg ubiquitination upon TG2 siRNA in p31e43-challenged T84cells. Immunoprecipitated PPARg species from whole-cell extracts of T84 cells immunoreactive for the anti-ubiquitin antibody were evident after challenge with p31e43 upon MG132 treatment. The simultaneous incubation with TG2 siRNA oligos induced a pronounced decrease of ubiquitinated PPARg protein. Immunoprecipitation (IP): anti-PPARg antibody; immunoblot (IB): anti-ubiquitin antibody. (C) Immunoblot analysis of PPARg protein in p31e43-challenged T84 cells. TG2 siRNA as well as cystamine inhibited p31e43-induced PPARg downregulation. Quantitative analysis (mean, SD) of three different experiments is reported in supplementary ﬁgure 3. (D) Confocal microscopy showed a decrease of p31e43- induced tyrosine phosphorylation in T84 cells upon rosiglitazone pre-treatment. (E) Phospho-tyrosine immunoreactivity of PY-99 immunoprecipitates of T84 cells after incubation with p31e43 upon TG2 siRNA in the presence or absence of GW9662. Decrease of PY99 immunoreactivity upon TG2 gene silencing. GW9662 inhibits TG2 siRNA-induced decrease of tyrosine phosphorylation. (F) Confocal microscopy shows an increase of p31e43-induced tyrosine phosphorylation in T84 cells upon TG2 gene silencing followed by GW9662. (D,E) Confocal microscopy, phospho-tyrosine (PY-99 antibody, green), DAPI (blue) nuclear counterstaining. Scale bar, 10 mm(AeF), are the results of three reproducible experiments. DAPI, 49-diamidine-29- phenylindol dihydrochloride; PPARg, peroxisome proliferator-activated receptor g; TG2, tissue transglutaminase.

CF TG2 drives the characteristic chronic inflammation via activated by dietary ligands,50 as well as commensal bacteria.51 In PPARg downregulation.12 this paper we provide the first evidence that TG2-mediated Here we demonstrate that in T84 cells, as well as in coeliac PPARg downregulation plays a key role in the pathogenesis of duodenum, p31e43 induces PPARg downregulation via ROS- coeliac disease. The effects of p31e43 are specific for coeliac mediated TG2. Therefore, an uncontrolled activation of the intestine and ‘gliadin-sensitive’ cell lines.30 31 10 This might be ROSeTG2 axis, either constitutive, as a consequence of genetic related to a coeliac disease-specific peptide internalisation or alterations as in CF,12 or induced by triggering factors such as reflect a still unknown coeliac disease-specific perturbation of the p31e43 in a susceptible target (coeliac mucosa and T84 epithelial machinery of peptide degradation. cells), leads to PPARg downregulation with a derangement of the The activation of the ROSeTG2 axis is therefore a key appropriate control of inflammation. PPARg is a hormone pathway of inflammation. TG2 works as a rheostat of ubiquiti- receptor produced by several cell types, including epithelial cells, nation and proteasome degradation in inflammation, particularly which negatively regulates inflammatory gene expression by in the context of coeliac disease. Indeed, in one case, increased ‘transrepressing’ inflammatory responses15 and even by modu- PPARg ubiquitination and degradation leads to increased lating oxidative stress.43 44 PPARg has been identified as a major inflammation, whilst increased TG2 ubiquitination and degra- functional receptor mediating the aminosalicylate activity45 46 dation leads to a reduced inflammatory response. A derangement and PPARg agonists have been exploited in therapeutic of TG2 regulation may lead to the amplification of the effects of approaches to control inflammation in chronic intestinal the stress. The induction of an epithelial pro-inflammatory inflammatory diseases such as ulcerative colitis47 48 and in phenotype may alter the first mucosal defence against ‘toxic’ experimental models of colitis.49 PPARg plays a key role in the agents and lead to a wide perturbation of the regulatory mech- regulation of the intestinal ‘inflammatory’ homeostasis since is anisms at the mucosal surface. The increased secretion of

Gut 2010;59:311e319. doi:10.1136/gut.2009.183608 317 Downloaded from gut.bmj.com on June 25, 2010 - Published by group.bmj.com

Coeliac disease

Figure 5 The challenge with p31e43 induces PPARg cross-linking and reduced PPARg protein in coeliac duodenum. (A) FRET analysis of celiac duodenal mucosa challenged for 3 h with p31e43. The increase of PPARgeAlexa-546 ﬂuorescence after Ne-(g-glutamyl)-L-lysine isopeptideeCy5 photobleaching in enterocytes of p31e43-challenged coeliac biopsies revealed PPARgeNe-(g-glutamyl)-L-lysine isopeptide interaction. (B,C) p31e43 challenge for 3 h induced a marked reduction of PPARg protein (green) in the enterocytes of duodenal mucosa from celiac patients (B) but not controls (C). The incubation with cystamine prevented p31e43 induced PPARg downregulation (B). (D) p31e43 challenge for 3 h induced high PY99 immunoreactivity (green) in the enterocytes of coeliac duodenal biopsies which was prevented by treatment with rosiglitazone. (BeD), Confocal microscopy, DAPI (blue) nuclear counterstaining. Scale bar, 10 mm. DAPI, 49-diamidine-29-phenylindol dihydrochloride; FRET, Fluorescence Resonance Energy Transfer; HM, high magniﬁcation; PPARg, peroxisome proliferator-activated receptor g.

inﬂammatory cytokines may, in turn, derange intestinal perme- Competing interests None. 52 ability and enhance the toxic effects of environmental triggers. Ethics approval This study was conducted with the approval of the Regione TG2 may be considered as a main player of the innate response to Campania Health Authority. stress inducers, thus setting the tone of whole mucosal response. Provenance and peer review Not commissioned; externally peer reviewed. Targeting the ROSeTG2 axis might represent a new pathogenic- based approach to antagonise the unwanted effects of gluten in coeliac disease. REFERENCES 1. Sollid LM. Coeliac disease: dissecting a complex inﬂammatory disorder. Nat Rev Acknowledgements The authors wish to thank Fabio Formiggini (Dynamic Imaging Immunol 2002;2:647e55. e Microscopy, CEINGE, Naples, Italy) for the technical support in FRET analysis. 2. Sollid LM. Molecular basis of coeliac disease. Annu Rev Immunol 2000;18:53 81. 3. Shan L, Molberg O, Parrot I, et al. Structural basis for gluten intolerance in celiac Funding This work was supported by the Coeliac UK and the Rothschild Trust sprue. Science 2002;297:2275e79. Corporation and Associazione Italiana Celiachia Regione Puglia (# 1400/07, Del. 4. Maiuri L, Ciacci C, Ricciardelli I, et al. Association between innate response to gliadin Reg.502 e 08/04/2008). and activation of pathogenic T cells in coeliac disease. Lancet 2003;362:30e7.

318 Gut 2010;59:311e319. doi:10.1136/gut.2009.183608 Downloaded from gut.bmj.com on June 25, 2010 - Published by group.bmj.com

Coeliac disease

5. Anderson RP, Degano P, Godkin AJ, et al. In vivo antigen challenge in celiac disease role of the innate immune response in Celiac disease. J Immunol 2006;176: identifies a single transglutaminase-modified peptide as the dominant A-gliadin T-cell 2512e21. epitope. Nat Med 2000;6:337e42. 29. Raki M, Tollefsen S, Molberg O, et al. A unique dendritic cell subset accumulates in 6. Gianfrani C, Levings MK, Sartirana C, et al. Gliadin-specific type 1 regulatory T cells the celiac lesion and efficiently activates gluten-reactive T cells. Gastroenterology from the intestinal mucosa of treated celiac patients inhibit pathogenic T cells. 2006;131:428e38. J Immunol 2006;177:4178e86. 30. Clemente MG, De VS, Kang JS, et al. Early effects of gliadin on enterocyte 7. Meresse B, Ripoche J, Heyman M, et al. Celiac disease: from oral tolerance to intracellular signalling involved in intestinal barrier function. Gut 2003;52:218e23. intestinal inflammation, autoimmunity and lymphomagenesis. Mucosal Immunol 31. Giovannini C, Matarrese P, Scazzocchio E, et al. Wheat gliadin induces apoptosis of 2009;2:8e23. intestinal cells via an autocrine mechanism involving FAS-FAS ligand pathway. FEBS 8. Meresse B, Chen Z, Ciszewski C, et al. Coordinated induction by IL15 of a TCR- Lett 2003;540:117e24. independent NKG2D signaling pathway converts CTL into lymphokine-activated killer 32. Jeyakumar M, Dwek RA, Butters TD, et al. Storage solutions: treating lysosomal cells in celiac disease. Immunity 2004;21:357e66. disorders of the brain. Nat Rev Neurosci 2005;6:713e25. 9. Barone MV, Gimigliano A, Castoria G, et al. Growth factor-like activity of gliadin, an 33. Verheijen FW, Verbeek E, Aula N, et al. A new gene, encoding an anion transporter, alimentary protein: implications for coeliac disease. Gut 2007;56:480e8. is mutated in sialic acid storage diseases. Nat Genet 1999;23:462e5. 10. Maiuri L, Ciacci C, Ricciardelli I, et al. Unexpected role of surface transglutaminase 34. Futerman AH, van Meer G. The cell biology of lysosomal storage disorders. Nat Rev type II in celiac disease. Gastroenterology 2005;129:1400e13. Mol Cell Biol 2004;5:554e65. 11. Esposito C, Marra M, Giuberti G, et al. Ubiquitination of tissue transglutaminase is 35. Terman A. Catabolic insufficiency and aging. Ann N Y Acad Sci 2006;1067: modulated by interferon alpha in human lung cancer cells. Biochem J 27e36. 2003;370:205e12. 36. Settembre C, Fraldi A, Jahreiss L, et al. A block of autophagy in lysosomal storage 12. Maiuri L, Luciani A, Giardino I, et al. Tissue Transglutaminase activation modulates disorders. Hum Mol Genet 2008;17:119e29. inflammation in Cystic Fibrosis via PPARg downregulation. J Immunol 37. Diez-Roux G, Ballabio A. Sulfatases and human disease. Annu Rev Genomics Hum 2008;180:7697e705. Genet 2005;6:355e79. 13. Schumann M, Richter JF, Wedell I, et al. Mechanisms of epithelial translocation of 38. Lorand L, Graham RM. Transglutaminases: crosslinking enzymes with pleiotropic the a2-gliadin-33mer in coeliac sprue. Gut 2008;57:747e54. functions. Nat Rev Mol Cell Biol 2003;4:140e56. 14. Matysiak-Budnik T, Candalh C, Dugave C, et al. Alterations of the intestinal 39. Juun E, Ronchetti RD, Quezado MM, et al. Tissue transglutaminase-induced transport and processing of gliadin peptides in celiac disease. Gastroenterology aggregation of alpha synuclein: implication for Lewy body formation in Parkinson’s 2003;125:696e707. disease and dementia with Lewy bodies. Proc Natl Acad Sci U S A 15. Daynes RA, Jones DC. Emerging roles of PPARs in inflammation and immunity. Nat 2003;100:2047e52. Rev Immunol 2002;2:748e59. 40. Zainelli GM, Ross CA, Troncoso JC, et al. Calmodulin regulates transglutaminase 2 16. Bailey ST, Ghosh S. ’PPAR’ting ways with inflammation. Nat Immunol cross-linking of hungtingtin. J Neurosci 2004;24:1954e61. 2005;6:966e7. 41. Verna A, Wang H, Manavathi B, et al. Increased expression of tissue 17. Chamberlain CG, Mansfield KJ, Cerra A. Glutathione and catalase suppress TGFb- transglutaminase in pancreatic ductal adenocarcinoma and its implications in drug induced cataract-related changes in cultured rat lenses and lens epithelial explants. resistance and metastasis. Cancer Res 2006;66:10525e33. Mol Vis 2009;15:895e905. 42. Bailey CD, Johnson GV. Tissue transglutaminase contributes to disease progression 18. Du X, Edelstein D, Obici S, et al. Insulin resistance reduces arterial prostacyclin in the R6/2 Huntington’s disease mouse model via aggregate-independent synthase and eNOS activities by increasing endothelial fatty acid oxidation. J Clin mechanisms. J Neurochem 2005;92:83e92. Invest 2006;116:1071e9. 43. Rizzo G, Fiorucci S. PPARs and other nuclear receptors in inflammation. Curr Opin 19. Vereb G, MatkoJ, Vamosi G, et al. Cholesterol-dependent clustering of IL-2Ralpha Pharmacol 2006;6:421e7. and its colocalization with HLA and CD48 on T lymphoma cells suggest their functional 44. Collino M, Aragno M, Mastrocola R, et al. Modulation of the oxidative stress and association with lipid rafts. Proc Natl Acad Sci U S A 2000;97:6013e18. inflammatory response by PPAR-gamma agonists in the hippocampus of rats exposed 20. Lu A, Tebar F, Alvarez-Moya B, et al. A clathrin-dependent pathway leads to KRas to cerebral ischemia/reperfusion. Eur J Pharmacol 2006;530:70e80. signaling on late endosomes en route to lysosomes. J Cell Biol 2009;6:863e79. 45. Liang HL, Ouyang Q. A clinical trial of combined use of rosiglitazone and 21. El Bekay R, Moises A, Javier M, et al. Oxidative stress is a critical mediator of the 5-aminosalicylate for ulcerative colitis. World J Gastroenterol 2008;14:114e19. angiotensin II signal in human neutrophils: involvement of mitogen-activated protein 46. Rousseaux C, Lefebvre B, Dubuquoy L, et al. Intestinal antiinflammatory effect of kinase, calcineurin, and the transcription factor NF-kB. Blood 2003;102:662e9. 5- aminosalicylic acid is dependent on peroxisome proliferator-activated receptor- 22. Andringa G, Lam KY, Chegary M, et al. Tissue transglutaminase catalyzes the gamma. J Exp Med 2005;201:1205e15. formation of a-synuclein cross-links in Parkinson’s disease. FASEB J 2004;18:932e4. 47. Dubuquoy L, Jansson EA, Deeb S, et al. Impaired expression of peroxisome 23. Karpuj MV, Becher MW, Springer JE, et al. Prolonged survival and decreased proliferatoractivated receptor gamma in ulcerative colitis. Gastroenterology abnormal movements in transgenic model of Huntington disease, with administration 2003;124:1265e76. of the transglutaminase inhibitor cystamine. Nat Med 2002;8:143e9. 48. Ramakers JD, Verstege MI, Thuijls G, et al. The PPARgamma agonist rosiglitazone 24. Maiuri L, Picarelli A, Boirivant M, et al. Definition of the initial immunologic impairs colonic inflammation in mice with experimental colitis. J Clin Immunol modifications upon in vitro gliadin challenge in the small intestine of celiac patients. 2007;27:275e83. Gastroenterology 1996;110:1368e78. 49. Bassaganya-Riera J, Reynolds K, Martino-Catt S, et al. Activation of PPAR gamma 25. Maiuri L, Ciacci C, Auricchio S, et al. Interleukin 15 mediates epithelial changes in and delta by conjugated linoleic acid mediates protection from experimental celiac disease. Gastroenterology 2000;119:996e1006. inflammatory bowel disease. Gastroenterology 2004;127:777e91. 26. Shidrawi RG, Day P, Przemioslo R, et al. In vitro toxicity of gluten peptides in coeliac 50. Marion-Letellier R, Dećhelotte P, Iacucci M, et al. Dietary modulation of peroxisome disease assessed by organ culture. Scand J Gastroenterol 1995;30:758e63. proliferator-activated receptor gamma. Gut 2009;58:586e93. 27. Palova Jelikova L, Rozkova D, Pecharova B, et al. Gliadin fragments induce 51. Dubuquoy L, Rousseaux C, Thuru X, et al. PPARgamma as a new therapeutic target in phenotypic and functional maturation of human dendritic cells. J Immunol inflammatory bowel diseases. Gut 2006;55:1341e9. 2005;175:7038e45. 52. Lammers KM, Lu R, Brownley J, et al. Gliadin induces an increase in intestinal 28. Thomas KE, Sapone A, Fasano A, et al. Gliadin stimulation of murine macrophage permeability and zonulin release by binding to the chemokine receptor CXCR3. inflammatory gene expression and intestinal permeability are MyD88-dependent: Gastroenterology 2008;135:194e204.

Gut 2010;59:311e319. doi:10.1136/gut.2009.183608 319 Downloaded from gut.bmj.com on June 25, 2010 - Published by group.bmj.com

PostScript

approved clinical trial and registry. However, Correspondence to Dr D K Bartsch, Department of The latest online pdf and full text have it is a fact and needs to be pointed out, that Visceral-, Thoracic and Vascular Surgery, Philipps been corrected. The journal apologises for the only one third of our identified IAR (80 of University Marburg, Germany; error. 205) participated in the recommended [email protected] screening programme. A pilot study on 32 of Competing interests None. doi:10.1136/gut.2008.155226corr1 these IAR using standard questionnaires and Provenance and peer review Not commissioned; not et al interviews (Beck Depression Inventory (BDI) De-Xin Zhang, Peng-Tao Zhao, Lin Xia, . externally peer reviewed. Gut 59 e and Brief Symptom Inventory (BSI)) around 2010; :292 9. Potent inhibition of e counselling (days À7, 0, +30) conducted by Gut 2010;59:1006 1007. doi:10.1136/gut.2010.214163 human gastric cancer by HER2-directed a psychiatrist revealed, that these IAR were induction of apoptosis with anti-HER2 critically biased by cognitive coping strategies REFERENCES antibody and caspase-3 fusion protein. Panels e fi (unpublished data). Pancreatic cancer (PC) 1. Harinck F, Canto MI, Schulick R, et al. Surveillance in E H were missing in gure 5. The latest screening is clearly different from other cancer individuals at high risk of pancreatic cancer; too early online pdf and full text have been corrected. e screening programmes, given the disastrous to tell? Gut 2010;59:1005 6. The journal apologises for this error. 2. Langer P, Kann PH, Fendrich V, et al. Five years of prognosis of PC, the unknown true pene- prospective screening of high-risk individuals from doi:10.1136/gut.2008.174904corr1 trance in the different settings of hereditary families with familial pancreatic cancer. Gut PC, the lack of a major gene defect, the lack of 2009;58:1410e18. Yue H-Y, Yin C, Hou J-L, et al. Gut 2010;59; reliable imaging or biomarkers, the lack of 3. Canto MI, Goggins M, Hruban RH, et al. Screening for 236e46. Hepatocyte nuclear factor 4a evidence to improve prognosis or to save lives early pancreatic neoplasia in high-risk individuals: attenuates hepatic fibrosis in rats. Professor by any screening, and the high risk of a prospective controlled study. Clin Gastroenterol Lin Yong and Dr Weifen Xie should have morbidity and mortality of potential preven- Hepatol 2006;4:766e81. quiz 665. been co-corresponding authors on this paper. tive surgery. Some authors even advocate that 4. Poley JW, Kluijt I, Gouma DJ, et al. The yield of first- The latest online pdf and full text have been at present ‘doing nothing’ provides the time endoscopic ultrasonography in screening individuals at a high risk of developing pancreatic corrected. greatest remaining quality of life-adjusted e 6 cancer. Am J Gastroenterol 2009;104:2175 81. years and the lowest costs. 5. Brand R, Rubinstein C, Lerch MM, et al. Advances in doi:10.1136/gut.2008.155853corr1 We fully agree that we need to gain much counselling and surveillance of patients at risk for more knowledge about hereditary PC to pancreatic cancer. Gut 2007;56:1460e9. Park DH, Kim M-H, Chari S T. Gut fi draw a de nite conclusion about the true 6. Rubenstein JH, Scheiman JM, Anderson MA. 2009;58:1680e9. Recent advances in auto- value of PC screening in IAR. However, A clinical and economic evaluation of endoscopic immune pancreatitis. M-H Kim and S T based on our data, we strongly believe, in ultrasound for patients at risk for familial pancreatic Chari should have been co-corresponding adenocarcinoma. Pancreatology 2007;7:514e25. accordance with the recommendations of authors on this paper. The latest online pdf the Fourth International Symposium of and full text have been corrected. Inherited Diseases of the Pancreas,5 that all CORRECTIONS screening procedures should be performed as doi:10.1136/gut.2009.179606corr1 part of peer-reviewed protocols combined doi:10.1136/gut.2009.208975ecorr1 with a scientific appraisal of the screening Mangia A, Andriulli A. Tailoring the length of methods and human subject protection. At Safety and comfort for colonoscopy in the antiviral treatment for Hepatitis C. Gut present there is no data, that would justify over seventies-time to achieve a balance? 2010;59:1e5. There were several errors in a general PC screening even of high risk Gut 2010;59(Suppl 1):A23. The correct table 1 which have now been corrected online. individuals outside of such protocols as author list should have been Hancock J, Ali F, suggested by Harinck et al.Incontrast,it Sarkar A, Parr J. doi:10.1136/gut.2008.174607corr1 has to be feared that uncritical use and Gut interpretation of screening results obtained doi:10.1136/gut.2009.190439corr1 Cahill RA, Lindsey I, Cunningham C. 58 e with the presently available tools on 2009; :1168 9. NOTES for colorectal Gut 59 d a healthcare basis may cause unnecessary Ibeakanma C, Vanner S. 2010; : neoplasia surgery through the looking glass. e physical harm and psychological distress. 612 21. TNFa is a key mediator of the pro- The surname of the third author should be On the other hand over-estimation of the nociceptive effects of mucosal supernatant spelt Cunningham, not Cunnigham. The latest power of our present screening tools may from human ulcerative colitis on colonic online pdf and full text have been corrected. fi lead to a deceptive, unjustified and poten- DRG neurons. The gures were ordered doi:10.1136/gut.2009.183608corr1 tially dangerous level of safety, if done incorrectly in the print version of the paper. uncritically and uncontrolled. The message The latest online pdf and full text have been Lysosomal accumulation of gliadin p31e43 of our paper thus is not ‘to do nothing’,but corrected. The journal apologises for the peptide induces oxidative stress and tissue to carefully evaluate screening methods for error. transglutaminase-mediated PPARg down- IAR from familial pancreatic cancer (FPC) regulation in intestinal epithelial cells and families in the setting of board approved doi:10.1136/gut.2009.180182corr1 coeliac mucosa. Gut 2010;59:311e9. The clinical trials, to continuously improve our Long-term outcome of endoscopic dilatation correct author list should have been Luciani knowledge and strategies. in patients with Crohn’s disease is not A, Rachela Villella V, Vasaturo A, Giardino I, affected by disease activity or medical Pettoello-Mantovani M, Guido S, Cexus O 1 2 1 P Langer, TM Gress, DK Bartsch therapy. Gut 2010;59:320e4. The correct N, Peake N, Londei M, Quaratino S, Maiuri L 1Department of General Surgery, Philipps-University author list should have been Thienpont C, as it appeared in the online first version. We Marburg, Germany; 2Department of Gastroenterology, D’Hoore A, Vermeire S, Demedts I, Bisschops have replaced the latest online pdf and full Philipps-University Marburg, Germany R, Coremans G, Rutgeerts P, Van Assche G. text. The journal apologises for the error.

Gut July 2010 Vol 59 No 7 1007 Published July 22, 2009, doi:10.4049/jimmunol.0900993 The Journal of Immunology

SUMOylation of Tissue Transglutaminase as Link between Oxidative Stress and Inﬂammation1

Alessandro Luciani,*† Valeria Rachela Villella,‡ Angela Vasaturo,‡ Ida Giardino,§ Valeria Raia,¶ Massimo Pettoello-Mantovani,* Maria D’Apolito,* Stefano Guido,†‡ Teresinha Leal,ʈ Sonia Quaratino,# and Luigi Maiuri2*,**

Cystic fibrosis (CF) is a monogenic disease caused by mutations in the CF transmembrane conductance regulator (CFTR) gene. CF is characterized by chronic bacterial lung infections and inflammation, and we have previously reported that tissue transglutaminase (TG2), a multifunctional enzyme critical to several diseases, is constitutively up-regulated in CF airways and drives chronic inflammation. Here, we demonstrate that the generation of an oxidative stress induced by CFTR-defective function leads to protein inhibitor of activated STAT (PIAS)y-mediated TG2 SUMOylation and inhibits TG2 ubiquitination and proteasome degradation, leading to sustained TG2 activation. This prevents peroxisome proliferator-activated receptor (PPAR)␥ and IkB␣ SUMOylation, leading to NF-␬B activation and to an uncontrolled inflammatory response. Cellular homeostasis can be restored by small ubiquitin-like modifier (SUMO)-1 or PIASy gene silencing, which induce TG2 ubiquitination and proteasome degradation, restore PPAR␥ SUMOylation, and prevent IkB␣ cross-linking and degradation, thus switching off inflammation. Manganese superoxide dismutase overexpression as well as the treatment with the synthetic superoxide dismutase mimetic EUK-134 control PIASy-TG2 interaction and TG2 SUMOylation. TG2 inhibition switches off inflammation in vitro as well as in vivo in a homozygous F508del-CFTR mouse model. Thus, TG2 may function as a link between oxidative stress and inflammation by driving the decision as to whether a protein should undergo SUMO-mediated regulation or degradation. Targeting TG2-SUMO interactions might represent a new option to control disease evolution in CF patients as well as in other chronic inflammatory diseases, neurodegenerative pathologies, and cancer. The Journal of Immunology, 2009, 183: 0000–0000.

ystic fibrosis (CF)3 is the most common life-threatening tissues (Online Mendelian Inheritance in Man no. 219700) (1, 2). inherited disease among the Caucasian population world- The most common CFTR mutation producing CF is deletion of C wide with considerable morbidity and reduced life ex- phenylalanine at residue 508 (F508del) in its amino acid sequence pectancy (1). The birth prevalence of CF is estimated to be in (1, 2). The misfolded F508del-CFTR protein is degraded and fails 3500–4500, with 200–300 new cases each year in single geo- to reach the cell membrane, leading to defective chloride channel graphic areas (1). CF is caused by mutations in the CF transmem- function (3–6). brane conductance regulator (CFTR) gene, which encodes a As a prototype of a monogenic disease, gene therapy is ideally cAMP-regulated chloride channel expressed at the apical mem- placed to treat CF patients, although this approach has so far not brane of epithelial cells in the airways, pancreas, testis, and other proved to be beneficial (7). Alternative therapeutic approaches have therefore been considered, such as small-molecule correctors of defective F508del-CFTR folding/cellular processing (“correc- *Institute of Pediatrics, University of Foggia, Foggia, Italy; †Dynamic Imaging Mi- croscopy, Centro di Ingegneria Genetica, Naples, Italy; ‡Department of Chemical tors”) and channel gatings (“potentiators”) (8, 9). Engineering, University Federico II of Naples, Naples, Italy; §Department of Labo- Although CF is a systemic disease, the main cause of death is ratory Medicine, University of Foggia, Foggia, Italy; ¶Department of Pediatrics, Uni- versity Federico II of Naples, Naples, Italy; ʈClinical Chemistry, Universite´ due to chronic airway inflammation and persistent and untreatable Catholique de Louvain, Brussels, Belgium; #Cancer Research UK Oncology Unit, pulmonary infections, with Pseudomonas aeruginosa colonizing University of Southampton, Southampton, United Kingdom; and **European Insti- most of the patients (10–12). As inflammation has a central role in tute for Research in Cystic Fibrosis, San Raffaele Scientific Institute, Milan, Italy the pathogenesis of CF (13), new antiinflammatory drugs provided Received for publication March 30, 2009. Accepted for publication June 21, 2009. some encouraging results in recent clinical trials. However, con- The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance flicting results have been reported on whether CF airways undergo with 18 U.S.C. Section 1734 solely to indicate this fact. proinflammatory response in the absence of bacterial infections 1 This work was supported by the European Institute for Research in Cystic Fi- (14–16). Several studies have shown that both cytokine profile brosis, Cancer Research UK, Rothschild Trust, Coeliac UK, and Regione Cam- secretion and NF-␬B activation are similar in CF and normal cells pania (L.229/99) and Associazione Italiana Celiachia Puglia (N.1400/07). (17, 18). However, mounting evidence suggests that inflammation 2 Address correspondence and reprint requests to Prof. Luigi Maiuri, European Insti- may occur before infection, and CFTR-defective cells have an in- tute for Research in Cystic Fibrosis, San Raffaele Scientific Institute, 20132 Milan, Italy. E-mail address: [email protected] trinsically proinflammatory phenotype (19–23). Recently, it has 3 Abbreviations used in this paper: CF, cystic fibrosis; CFTR, CF transmembrane been shown that CFTR is a negative regulator of NF-␬B-mediated conductance regulator; F508del, deletion of phenylalanine at residue 508; TG2, tissue innate immune response, and its localization to lipid rafts is in- transglutaminase; PPAR, peroxisome proliferator-activate receptor; ROS, reactive oxygen species; SUMO, small ubiquitin-like modifier; PIAS, protein inhibitor of acti- volved in control of inflammation (24). vated STAT; NEMO, NF-␬B essential modulator; siRNA, small interfering RNA; We have previously investigated the molecular mechanisms that MnSOD, manganese superoxide dismutase; FRET, fluorescence resonance energy regulate this “intrinsic” proinflammatory profile and made a link transfer; N-CoR, nuclear co-repressor. between CFTR-defective function and inflammation (22). We have Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 shown that nasal polyp mucosa from CF patients as well as human

www.jimmunol.org/cgi/doi/10.4049/jimmunol.0900993 2 TG2 LINKS OXIDATIVE STRESS AND INFLAMMATION

CFTR-defective cell lines constitutively up-regulate tissue trans- Sigma-Aldrich) or buthionine sulfoximine (10 ␮M; Sigma-Aldrich) in the glutaminase (TG2), a multifunctional protein expressed in several presence or absence of EUK-134 as well as with CFTR-inh172 (20, 22) (20 ␮ tissues (25). The increased TG2 activity drives inflammation M; Calbiochem). A549 and 16HBE cell lines were also cultured under hypoxic conditions in a humidified hypoxia chamber (Cop Laboratories) through down-regulation of the antiinflammatory peroxisome pro- for1hat1%O2 (5% CO2 balance N2), and temperature was maintained liferator-activated receptor (PPAR)␥, a negative regulator of in- at 37°C (34). flammatory gene expression (26). RNA interference TG2 is an enzyme with a vast array of biological functions (27). It catalyzes cross-links or deamidation of target proteins in the IB3-1 cells were transfected with 50 nM human SUMO-1 or scrambled small presence of high Ca2ϩ levels (27), whereas at low Ca2ϩ concen- interfering RNA (siRNA) duplex using Lipofectamine RNAiMAX at 37°C for 72 h. The SUMO-1 duplex siRNA was a pool of three se- trations it may function as a G protein or a protein disulfide quences: siRNA no. 1, SUMO-1, sense, 5Ј-GGAUAGCAGUGAGAUU isomerase (27), thus contributing to the functionality of the mitho- CACUUCAAA-3Ј, antisense, 5Ј-UUUGAAGUGAAUCUCACUGCUA condrial respiratory chain complexes. We have shown that in CF UCC-3; siRNA no. 2, SUMO-1, sense, 5Ј-GGUGUUCCAAUGAAUU airways high levels of reactive oxygen species (ROS) lead to an CACUCAGGU-3Ј, antisense, 5Ј-ACCUGAGUGAAUUCAUUGGAAC Ј Ј ␥ ACC-3 ; siRNA no. 3, SUMO-1, sense, 5 -GGAAGAAGAUGUGAUU increase of TG2 activity, TG2-mediated PPAR cross-linking, GAAGUUUAU-3Ј, antisense 5Ј-AUAAACUUCAAUCACAUCUUCU ubiquitination, and proteasome degradation, thus driving inflam- UCC-3Ј. siRNA-mediated knockdown of PIASy was performed using mation (22). Blocking TG2 through specific gene-silencing or TG2 specific siRNA oligos, as previously described (33). TG2 gene silencing inhibitors increases PPAR␥ protein and reverses inflammation was performed as previously reported (22). (22). The mechanisms by which the genetic CF defect leads to Adenoviral vector TG2 up-regulation and inflammation, however, are still unknown. Human manganese superoxide dismutase (MnSOD) cDNA was cloned into TG2 is regulated by retinoids, steroid hormones, peptide growth the shuttle vector pAd5CMVK-NpA (35). MnSOD adenovirus was a gift factors, and cytokines that also lead to a time-dependent decrease from Michael Brownlee (Albert Einstein College of Medicine, New York, in TG2 ubiquitination (25). Increased TG2 tissue levels have been NY). IB3-1, A549, and 16HBE cell lines were infected with MnSOD or observed in neurodegenerative diseases (28), including Alzhei- control adenovirus for 2 h, as previously described (35). mer’s, Huntington’s, and Parkinson’s diseases, as well as in TG2 overexpression chronic inflammatory conditions such as celiac disease (29). In- A549 cells were transfected with wild-type TG2 cloned into the pLPCX creased tissue levels of TG2 have also been detected in cancer such vector (pLPCX-TG2) (a gift from Dr. G. M. Fimia, National Institute for as glioblastomas, malignant melanomas, and pancreatic ductal ad- Infectious Diseases, Istituto di Ricovero e Cura a Carattere Scientifico,“L. enocarcinomas (30), and they are often associated with an in- Spallanzani”, Rome, Italy) and empty vector (pLPCX), as control, using creased metastatic activity or acquisition of drug resistance (30). Lipofectamine 2000 (Invitrogen) according to the manufacturer’s protocol, In this paper we investigated whether posttranslational modifi- and incubated at 37°C for 48 h. cations result in the persistence of high levels of TG2 protein in CF Cell fractionation airways. We focused on small ubiquitin like-modifier (SUMO) IB3-1 cells were collected in cold buffer A (20 mM Tris-HCl (pH 7.4), 2 posttranslational modification, since this has been defined as a cen- mM EDTA, 20 mM 2-ME, 1ϫ PMSF, 1 ␮g/ml inhibitor protease cocktail), tral way of regulating key cellular functions and stability of pro- homogenized in Potter-Elvehjem pestle and glass tube (Sigma-Aldrich), teins (31, 32). Since protein inhibitor of activated STAT (PIAS)y, and centrifuged at 2000 rpm for 15 min at 4°C to obtain nuclear pellets. a member of the PIAS family (33), has recently been defined as the Supernatants were collected as cytoplasmic fractions. Nuclear pellets were washed with buffer A and resuspended in buffer B (2 mM Na VO , 400 first SUMO ligase for NF-␬B essential modulator (NEMO) (33), 3 4 mM NaCl, 1 mM MgCl2, 1 mM EGTA, HEPES 10 mM (pH 7.9), 1 mM and PIASy-NEMO interaction is mediated by ROS (33), we also DTT, 1ϫ PMSF, 1 ␮g/ml inhibitor protease cocktail) and incubated on ice investigated whether PIASy-TG2 interaction could mediate ROS- for 50 min with occasional mixing to extract nuclear proteins. Nuclear driven posttranslational modifications of TG2. We demonstrate extracts were cleared by centrifugation (7000 rpm, 15 min, 4°C), and supernatants were collected as nuclear fraction. Then, cytoplasmic and nu- that in CF airway epithelia the pro-oxidative intracellular milieu clear whole-cell fractions were analyzed by immunoblotting. leads to PIASy-mediated TG2 SUMOylation and uncontrolled TG2 activation. This prevents PPAR␥ and IkB␣ SUMOylation, Immunoblot leading to NF-␬B activation and persistent inflammation. More- The blots were incubated with anti-phospho-p42/p44 MAP kinases (rabbit over SUMO-1 or PIASy gene silencing can switch off inflamma- polyclonal IgG, 1/500; Cell Signaling Technology), PPAR␥ (clone E8 sc- tion, favoring TG2 degradation and restoring cellular homeostasis. 7273, mouse monoclonal IgG1, 1/500; Santa Cruz Biotechnology), TG2 (clone CUB7402, mouse monoclonal IgG1, 1/500; NeoMarkers), ubiquitin We demonstrate increased ROS levels and TG2 SUMOylation (clone FL-76 sc9133, rabbit polyclonal IgG, 1/500; Santa Cruz Biotech- also in vivo, in a mouse model homozygous for the F508del-CFTR nology), SUMO-1 (clone FL-101 sc9060, 1/500; Santa Cruz Biotechnol- mutation, and show that TG2 inhibition blocks lung inflammation ogy), phopsho-p65(Ser536) (rabbit polyclonal, 1/500; Cell Signaling Tech- in vivo. nology), PIASy (clone H75 sc50437, rabbit polyclonal IgG, 1/1000; Santa ␬ ␣ These results indicate TG2 as a missing link between oxidative Cruz Biotechnology), I B (clone H4 sc1643, mouse monoclonal IgG1, 1/1000; Santa Cruz Biotechnology), ␤-actin (clone 13E5, rabbit polyclonal stress and inflammation and flag TG2 as a new attractive target to IgG, 1/2000; Cell Signaling Technology), and ␣␤-tubulin (rabbit poly- control the evolution of chronic airway inflammation in CF. clonal IgG, 1/2000; Cell Signaling Technology). The primary Abs were counterstained by a HRP-conjugated anti-IgG Ab (Amersham Biosciences) Materials and Methods for 60 min at room temperature. Proteins were visualized by chemiluminescence (ECL Plus; Amersham Biosciences) and exposed to X-OMAT Cell lines and cultures film (Eastman Kodak). The amounts of proteins were determined by a Human CF bronchial epithelial cell line IB3-1, carrying F508del/W1282X Bio-Rad protein assay to ensure equal protein loading before Western blot CFTR mutation, isogenic stably rescued C38, normal bronchial epithelial analysis. Fifty micrograms of cell lysate was loaded in each lane. 16HBE, or lung carcinoma A549 cell lines (LGC Standards) were cultured Immunoprecipitation as recommended by American Type Culture Collection. IB3-1 cell lines were incubated for 24 h with EUK-134 (50 ␮g/ml; Alexis Biochemicals), Treated and untreated cells were harvested, lysed, and 500 ␮g of cell lysate TG inhibitor R283 (250 ␮M), KCC009 (250 ␮M), or cystamine (400 ␮M; was immunoprecipitated by overnight incubation at 4°C on a mixer with an Sigma-Aldrich) followed or not by rosiglitazone for6h(10␮M; Alexis appropriate dilution of specific Ab (anti-TG2 CUB 7402 mAb, anti-PIASy, Biochemicals). A549 and 16HBE cell lines were incubated for 1 h with anti-PPAR␥, anti-I␬B␣) in cold lysis buffer. The samples were then incu- ␮ H2O2 (33) (2 mM; Sigma-Aldrich) and for 24 h with rotenone (1 M; bated with protein G-Sepharose at 4°C for 2 h with constant mixing. After The Journal of Immunology 3

FIGURE 1. TG2 SUMOylation in CF airway epithelial cells. A, Immu- noblot analysis of SUMO-1 expression in CF IB3-1 and C38 cells. ␤-actin levels were used as loading control. B, FRET analysis of SUMO-1-TG2 interaction in IB3-1 and C38 cells. C, Immunoprecipitated (IP) TG2 species from whole-cell extracts of IB3-1 cells are immunoreactive for the anti-SUMO-1 Ab. D, Immunoblot analysis of TG2 immunoprecipitates FIGURE 2. PIASy mediates TG2 SUMOylation in CF airway epithelial with anti-TG2 Abs. Two TG2 bands are detected in IB3-1 cells. E, Immu- cells. A–C, IB3-1 cells were transduced with either human MnSOD or noblot analysis of TG2 expression in CF IB3-1 and C38 cells. ␤-actin antisense cDNAs in pAd5CMVK vector. A, Immunoblot analysis of TG2 levels were used as loading control. (left) or PIASy (right) immunoprecipitates (IP) with anti-PIASy (left)or anti-TG2 (right) Abs, respectively. B, Immunoblot analysis with anti- SUMO-1 Ab of TG2 immunoprecipitates. C, Immunoblot analysis of washing, the immunoprecipitated proteins were electrophoresed through SUMO-1, PIASy, and TG2 protein. ␤-actin levels were used as loading 10% polyacrylamide gels (Bio-Rad), transferred onto blotting membranes control. D, IB3-1 cells were transfected with either 50 nM human PIASy (PolyScreen polyvinylidene difluoride; NEN), and analyzed. siRNA or control siRNA. FRET analysis of SUMO-1-TG2 interaction. Patients and ex vivo cultures of nasal polyp mucosal biopsies Seven consecutive CF patients carrying the common CFTR mutations (F508del/F508del, F508del/W1282X, F508del/N1303K, or F508del/ ter frozen lung tissue sections from each mouse were pulsed with 10 ␮M G542X) (mean age, 19 years; range, 13–29 years) with chronic sinusitis CM-H2DCFDA and analyzed by confocal microscopy. and nasal polyposis and seven consecutive non-CF patients (mean age, 21 years; range, 16–32 years) with idiopathic polyposis underwent surgical Confocal microscopy treatment. Informed consent was obtained from all subjects and the ethical Cell lines. Treated or untreated cells were fixed in methanol, permeabil- committee of Regione Campania Health Authority approved the study. CF ized with 0.5% Triton X-100, and incubated with Abs against PPAR␥ nasal polyp biopsies were cultured for 4 h, as previously reported (22), with (1/100 dilution; Santa Cruz Biotechnology), SUMO-1 (1/100; Santa Cruz or without ROS scavenger EUK 134 (50 ␮g/ml). Control nasal polyp bi- Biotechnology), TG2 (1/100; NeoMarkers), PIASy (1/100; Santa Cruz opsies were also cultured for 1 h with H2O2 (2 mM; Sigma-Aldrich). Biotechnology), and nuclear co-repressor (N-CoR, 1/200; Santa Cruz Bio- technology) according to the previously described procedure (22). Mice Human tissue sections. Five-micrometer frozen human lung tissue sec- Young adult female CF mice homozygous for the F508del mutation in tions were fixed in acetone for 10 min. The sections were incubated for 2 h the 129/FVB outbred background (36) and their wild-type littermates were at room temperature with Abs against SUMO-1 (1/100; Santa Cruz Bio- housed in static isolator cages at the animal care specific pathogen-free technology), PIASy (1/100; Santa Cruz Biotechnology), and TG2 (1/100; facility of the University of Louvain following recommendations of the NeoMarkers) as previously described (22). Federation of European Laboratory Animal Science Associations (37). To Mice lung tissue. Seven-micrometer frozen lung tissue sections from each prevent intestinal obstruction CF mice were weaned to a liquid diet (Pep- mice were fixed in acetone for 10 min. The sections were incubated for 2 h tamen; NestleŃutrition). Peptamen was replaced daily. The genotype of at room temperature with the following Abs: anti-phospho-p42/p44 MAP each animal was checked at 21 days of age, as previously described (36). kinases (1/200; Cell Signaling Technology), PPAR␥ (clone H100 sc-7196, These studies and procedures were approved by the local Ethics Committee rabbit polyclonal IgG, 1/100; Santa Cruz Biotechnology), TG2 (clone for Animal Welfare and conformed to the European Community regula- H237 sc20261, rabbit polyclonal IgG, 1/100; Santa Cruz Biotechnology), tions for animal use in research (CEE no. 86/609). CF (n ϭ 7) and normal and SUMO-1 (clone S-18 sc31852, goat polyclonal IgG, 1/100). This was homozygous wild-type mice (n ϭ 7), 10 to 14 wk of age, were treated i.p. followed by incubation with Alexa 488 donkey anti-rabbit (used for de- (38) for 7 days with a daily dose of 100 ␮l of 0.01 M cystamine or PBS tection of phospho-p42–44, PPAR␥, and TG2 protein, 1/200; Invitrogen) solution. Mice were then killed by i.p. injection of 20 mg of sodium pen- or 546 donkey anti-goat (used for detection of SUMO-1 protein, 1/200; tobarbital (Abbott Laboratories). Invitrogen). Data were analyzed under fluorescence examination by confocal microscopy as previously described (22). ROS detection In situ detection of TG2 enzymatic activity Cell lines were pulsed with 10 ␮M 5-(and-6)-chloromethyl-2Ј7Ј-dichlorodihydrofluorescein diacetate acetyl ester (CM-H2DCFDA) (Molecular TG2 activity in cell lines was detected by incubating live cells with bio- Probes/Invitrogen) according to the manufacturer’s suggestions. The cells tinylated monodansylcadaverine for1hat37°C (16). The incorporation of were analyzed with a Wallac 1420 multilabel counter (PerkinElmer) or labeled substrate was visualized by incubation with Alexa 546-conjugated detected by a LSM510 Zeiss confocal laser scanning unit. Seven-microme- streptavidin (1/100; Molecular Probes/Invitrogen) for 30 min (22). TG2 4 TG2 LINKS OXIDATIVE STRESS AND INFLAMMATION

FIGURE 3. Decrease of TG2 protein levels and TG2 activity by either SUMO or PIASy gene silencing. A–C, IB3-1 cells were transfected with either 50 nM human SUMO-1 siRNA or control siRNA. A, Im- munoblot analysis with anti-ubiquitin Ab of TG2 immunoprecipitates upon SUMO-1 gene silencing followed by a 6-h incubation with 50 ␮M proteasome inhibitor MG132. B, Immunoblot analysis of SUMO-1 and TG2 protein in IB3-1 cells upon SUMO-1 gene silencing. ␤-actin levels were used as loading control. C, TG2 activity (white) in IB3-1 cells upon SUMO-1 gene silencing. Confocal microscopy, scale bar, 10 ␮m. D–F, IB3-1 cells were transfected with either 50 nM human PIASy siRNA or control siRNA. D, Immunoblot analysis with anti-ubiquitin Ab of TG2 immunoprecipitates upon PIASy gene silencing followed by a 6-h incubation with 50 ␮M proteasome inhibitor MG132. E, Immu- noblot analysis of PIASy and TG2 protein upon PIASy gene silencing. ␤-actin levels were used as loading control. F, TG2 activity (white) in IB3-1 cells upon PIASy gene silencing. Confocal microscopy, scale bar, 10 ␮m.

enzymatic activity on human or mice lung sections was detected by FIGURE 4. ROS-mediated PIASy-TG2 interaction in human nasal incubating unfixed sections with biotinylated monodansylcadaverine for A ␥ 1 h at room temperature, as previously described (22). polyp mucosa. , Confocal images of PIAS (green) and TG2 protein (red) in CF nasal polyp mucosa. The same pattern is observed after a 4-h incu- Fluorescence resonance energy transfer (FRET) microscopy bation with medium alone. B, Confocal images of PIAS␥ and TG2 in CF nasal polyp mucosa cultured for 4 h with EUK-134. C, Confocal images of For acceptor photobleaching, cells were fixed with buffered 2% paraformald- PIAS␥ (green) and TG2 protein (red) in control nasal polyp mucosa. D, heyde and permeabilized with 0.5% Triton X-100. Upon fixation, cells were ␥ immunostained with Alexa 546-anti-TG2 (Molecular Probes/Invitrogen)/ Confocal images of PIAS and TG2 in control nasal polyp mucosa cul- Cy5-anti-SUMO-1 (Santa Cruz Biotechnology) or Alexa 546-anti-TG2/ tured for1hwith2mMH2O2. A–D, Overlay shows the merged images. Cy5-anti-PIASy (Santa Cruz Biotechnology). Seven-micrometer frozen DAPI (4,6-diamidino-2-phenylindole) (blue), nuclear counterstaining. lung tissue sections from each mouse were fixed with buffered 2% Scale bar, 10 ␮m. paraformaldheyde. Upon fixation, tissue sections were immunostained with Alexa 546-anti-TG2/Cy5-anti-SUMO-1 (Santa Cruz Biotechnol- ogy). Cells or tissue sections were then mounted for performing FRET parisons were evaluated by one-way ANOVA, and post hoc compari- assay by confocal microscopy. Cy5 was bleached at Յ10% of its initial sons were made using ANOVA Tukey-Kramer test. A p value of Ͻ0.05 fluorescence, by 200 pulses (2.56 ms) of 5 mW 633 nm laser per pixel, was considered significant. sampling 0.01 mm2 of the specimen. Alexa 546 fluorescence was detected before and after Cy5 photobleaching (39). Results ELISA PIASy-mediated TG2 SUMOylation increases TG2 protein levels in CF airway epithelial cells Human or murine TNF-␣ secretion was measured using the BD OptEIA TNF-␣ ELISA kit II (BD Biosciences). Measurements were performed at To investigate whether posttranslational modifications could result least in triplicate. Values were normalized to 106 cells; results were ex- in the persistence of high levels of TG2 protein, we first analyzed Ϯ pressed as means SEM. IB3-1 CF epithelial cell lines carrying F508-del/W1282X CFTR Statistical analysis mutations and the isogenic stably rescued C38 cells. Western blots revealed that SUMO-1 was increased in IB3-1 cells as compared All experiments were performed at least in triplicate. Data distribution was analyzed, and statistical differences were evaluated by using with C38 cell lines (Fig. 1A). Furthermore, acceptor photobleach- ANOVA Tukey-Kramer test by SPSS 12 software. Results of experi- ing FRET revealed that SUMO-1 interacted with TG2 in IB3-1 ments on mice were expressed as means Ϯ SEM. Between-group com- cells (Fig. 1B), and SUMO-1 immunoreactivity was detected on The Journal of Immunology 5

FIGURE 5. Deregulation of ROS machinery mediates TG2 SUMOyla- tion in 16HBE and A549 cells. A and B, 16HBE cell lines were cultured with or without CFTRinh-172 in the presence or absence of MnSOD overexpression. A, FRET analysis reveals SUMO-1- TG2 interaction after CFTR inhibition that was controlled by MnSOD overexpression. B, Immunoblot analysis of PIASy protein. ␤-actin levels were used as loading control. C and D, A549 cells were cultured with or without rotenone. C, Immunoprecipitated (IP) TG2 species from whole-cell extracts are immunoreactive for the anti-SUMO-1 Ab upon rotenone treatment. D, FRET analysis reveals SUMO-1-TG2 interaction after incubation with rotenone. MnSOD overexpression controls rotenone-induced TG2 SUMOylation. E–G, A549 cells were cultured with

H2O2 or under hypoxic conditions. E, Immunoblot analysis of PIASy protein ␤ after H2O2 treatment. -actin levels were used as loading control. F, Immu- noblot analysis of TG2 immunoprecipitates (IP) with anti-PIASy Ab after

H2O2 treatment in presence or absence of MnSOD overexpression. G, FRET analysis reveals SUMO-1-TG2 interaction after H2O2 treatment as well as when A549 cells were cultured upon hypoxic conditions. The effects of hypoxia on TG2 SUMOylation are controlled by MnSOD overexpression.

TG2 immunoprecipitates (Fig. 1C). When TG2 immunoprecipi- reducing TG2 protein levels (Fig. 2C). Furthermore, MnSOD over- tates from IB3-1 cells were probed with the anti-TG2 Ab, two TG2 expression reduced PIASy protein levels (Fig. 2C). The antioxi- bands were detected, with the upper band corresponding to the dant synthetic SOD mimetic EUK-134 (22) showed the same SUMOylated TG2 (Fig. 1D). Moreover, TG2 protein levels were effects as MnSOD (supplemental Fig. 2). We confirmed the in- higher in IB3-1 than in C38 cell lysates (Fig. 1E). Sequence anal- volvement of PIASy in TG2 SUMOylation by PIASy gene silenc- ysis and SUMO motif screening revealed three SUMO acceptor ing. We reduced PIASy cellular expression by Ͼ90%, using sites (␺_kxE) (32) in TG2 sequence (supplemental Fig. 1).4 PIASy siRNA. Indeed, PIASy siRNA inhibited TG2-SUMO-1 We investigated whether PIASy could mediate TG2 SUMO- interaction (Fig. 2D) and TG2-SUMO-1 coimmunoprecipitation ylation. We found that PIASy and TG2 coimmunoprecipitated in (data not shown). This suggests that the increase of ROS levels IB3-1 cells (Fig. 2A). Since PIASy SUMO-1 E3-ligase activity is induces PIASy-TG2 interaction and TG2 SUMOylation. influenced by ROS (33), we investigated whether the CF intracel- SUMOylation may induce protein stabilization by blocking lular prooxidative environment (22) could mediated TG2-PIASy ubiquitination of the same lysine residues (40). We demonstrated interaction and TG2 SUMOylation. The overexpression of human that SUMO gene silencing by SUMO-1-specific siRNA increased MnSOD (35) controlled PIASy-TG2 coimmunoprecipitation (Fig. TG2 ubiquitination upon proteasome inhibition by MG132 (Fig. 2A) as well as SUMO-TG2 coimmunoprecipitation (Fig. 2B), thus 3A), thus allowing TG2 to be targeted to proteasome for degradation. This induced decreases of TG2 protein (Fig. 3B) and TG2 4 The online version of this article contains supplemental material. activity (Fig. 3C). PIASy siRNA showed the same effects as 6 TG2 LINKS OXIDATIVE STRESS AND INFLAMMATION

SUMO-1 siRNA on TG2 ubiquitination (Fig. 3D) and TG2 protein levels (Fig. 3E) and activity (Fig. 3F). To investigate whether TG2-PIASy interaction and TG2 SUMO- ylation may take place in human airways of CF patients and whether it was induced by the oxidative stress, we used a well- established tissue culture model of biopsies of human CF nasal polyps (22, 41). We have already validated this experimental model (22, 41) and reported that increased TG2 levels are a feature of CF nasal polyp mucosa and that the inhibition of TG2 is effective in controlling mucosal inﬂammation by restoring normal levels of PPAR␥ protein (22). We found that TG2-PIASy colocalized in human CF airways (Fig. 4A) and that this interaction was inhibited upon treatment with the EUK-134 (Fig. 4B). EUK-134 reduced TG2 expression at epithelial but not in the subepithelial compartment (Fig. 4B, arrow), where PIASy did not colocalize with TG2 protein (arrow in Fig. 4A). After EUK-134 treatment the distribution of TG2 in CF nasal polyp biopsies was similar to that observed in non-CF controls (22). This suggests that the inhibition of TG2-PIASy interaction restores the physiological levels and distribution of TG2 in CF airways (22). Furthermore, in CF nasal polyp mucosa TG2 colocalizes with SUMO-1, and the incubation FIGURE 6. SUMO-1 or PIASy gene silencing control inﬂammation in with EUK-134 inhibited TG2-SUMO-1 colocalization (data not CF airway epithelial cells. IB3-1 cells were transfected with either 50 nM shown). Non-CF control nasal polyp mucosa showed very faint human SUMO-1 siRNA, PIASy siRNA, or control siRNA. A, Immunoblot analysis of p42–44 phosphorylation upon PIASy siRNA (left) and SUMO-1 TG2, PIASy, or SUMO-1 expression at the epithelial level (Fig. 4, siRNA (right). ␤-actin levels were used as loading control. B, TNF-␣ protein upon C and D). In nasal control biopsies the treatment with H2O2 was SUMO-1 or PIASy gene silencing (each bar represents the mean plus SEM of highly effective in increasing epithelial SUMO-1 and TG2 expres- .(p Ͻ 0.008 vs control siRNA). sion and their colocalization (Fig. 4D ,ء ;three separate experiments, each with n ϭ 3

FIGURE 7. TG2 inhibition modulates PPAR␥ and IkB␣ pathways in CF airway epithelia. A and B, IB3-I cells were incubated for 6 h with rosiglitazone in presence or absence of TG2 gene silencing. A, Im- munoprecipitated (IP) PPAR␥ from whole- cell extracts are immunoreactive for the anti-SUMO-1 Ab upon TG2 siRNA. B, Confocal images of IB3-1 cells immunostained with N-CoR (green). DAPI (4,6- diamidino-2-phenylindole) nuclear counterstaining. Scale bar, 10 ␮m. C and D, IB3-1 cells were transfected with either 50 nM human TG2 siRNA or control siRNA. C, Immunoprecipitated IkB␣ species from whole-cell extracts are immunoreactive for the anti-SUMO-1 Ab upon TG2 siRNA. D, Immunoblot analysis of IkB␣ expression upon TG2 gene silencing. E, Immunoblot analysis of phospho-p65(Ser536) in cytoplasmic (C) and nuclear (N) cell fractions upon TG2 gene silencing. F, IB3-1 cells were incubated with rosiglitazone in the presence or absence of 400 ␮M cystamine. Increased immunoreactivity of immunoprecipitated PPAR␥ species from whole-cell extracts for the anti-SUMO-1 Ab after cystamine treatment is shown. G, Increased immunoreactivity of immunoprecipitated IkB␣ species from whole-cell extracts of IB3-1 cells for the anti- SUMO-1 Ab after cystamine treatment. H and I, IB3-1 cells were transfected with wild- type TG2 cloned into the pLPCX vector (pLPCX-TG2) or empty vector (pLPCX). Decrease of SUMO-1 immunoreactivity of immunoprecipitated PPAR␥ (H)orIkB␣ (I) species from whole-cell extracts after TG2 overexpression. The Journal of Immunology 7

Deregulation of ROS machinery mediates TG2 SUMOylation in human airway epithelial cell lines Since we have previously demonstrated that the inhibition of CFTR function by CFTR-172inh leads to increases of ROS levels and TG2 activation (22), we investigated whether a deregulation of ROS machinery could mediate TG2 SUMOylation in airway epithelia. We inhibited CFTR function by CFTR-172inh in 16HBE cell lines and demonstrated that PIASy-TG2 interaction (data not shown) and TG2 SUMOylation (Fig. 5A) were induced upon CFTR inhibition, which up-regulates intracellular ROS (22). MnSOD overexpression was highly effective in controlling CFTR- 172inh-induced TG2 SUMOylation (Fig. 5A). The same effects were observed in A549 epithelial cell lines (data not shown). CFTR-172inh also increased PIASy protein levels (Fig. 5B). Moreover, rotenone, the most commonly used complex I inhibitor that increases ROS production in mitochondria (42), induced TG2- SUMO-1 coimmunoprecipitation (Fig. 5C) and interaction (Fig. 5D) in A549 cells. The effects of rotenone (1 ␮M) were neutralized by MnSOD overexpression (Fig. 5D) as well as by EUK-134 (data not shown). Buthionine sulfoximine, an inhibitor of the glutathione pathway (42), showed the same effects as those observed after treatment with rotenone (supplemental Fig. 2).

The treatment of A549 cells with H2O2 induced increased PIASy protein levels (Fig. 5E) as well as PIASy-TG2 interaction (Fig. 5F) and TG2 SUMOylation (Fig. 5G). These effects were controlled by MnSOD overexpression (Fig. 5F). Moreover, when A549 cells were cultured under hypoxic conditions (34), increased TG2 SUMOylation was observed (Fig. 5G). The effects of hypoxia were likely mediated by ROS (43) since MnSOD overexpression FIGURE 8. TG2 inhibition by cystamine controls airway inflammation controlled hypoxia-induced TG2 SUMOylation (Fig. 5G). in F508del-CFTR homozygous mice. A, Increase of intracellular ROS These data suggest that TG2 SUMOylation occurs as a conse- (DCF immunofluorescence) in F508del-CFTR mice as compared with quence of the deregulation of the oxidative control machinery. wild-type (WT) littermates. Confocal microscopy. Scale bar, 10 ␮m. B, FRET analysis of TG2-SUMO-1 interaction in homozygous F508del- SUMO-1 or PIASy gene silencing controls inflammation in CF CFTR mice and WT littermates. C–E, Homozygous F508del-CFTR mice airway epithelial cells treated with cystamine (daily i.p. injection of 100 ␮l of 0.01 M in PBS/7 days) or PBS. The pattern observed after PBS treatment was similar to that TG2 SUMOylation might provide the missing link between cel- observed in untreated mice. C, Confocal images identified increased TG2 lular stress and inflammation. We tested whether the control of activity, low expression of PPAR␥ protein with localization in perinuclear TG2 SUMOylation might modulate TG2-driven inflammation we aggregates, and increased p42–44 phopshorylation in lung tissues of PBS- have described in CF epithelia (22). We demonstrated that gene treated CF mice as compared with control littermates. Cystamine reduces silencing of either PIASy or SUMO by specific siRNAs induced TG2 activity and phospho-p42–44 and restores high epithelial expression a significant decrease of p42–44 phosphorylation (Fig. 6A) and and nuclear localization of PPAR␥ as observed in WT mice. D, High TNF-␣ release (Fig. 6B) in IB3-1 cells. magnification of C. Confocal microscopy, CyTRAK Orange nuclear counterstaining, yellow indicates nuclear localization. One representative case TG2 inhibition modulates PPAR␥ and IkB␣ SUMOylation in of seven mice for each group is shown. E, TNF-␣ protein in mouse lung CF airway epithelia homogenates (mean plus SEM of three separate experiments, each with .(p Ͻ 0.017 vs PBS-treated mice ,ء ;n ϭ 7 Since PPAR␥ may undergo SUMOylation in response to a PPAR␥ agonist (22, 44), thus interacting with N-CoR-histone deacetylase 3 (HDAC3) complex and thereby blocks its ubiquitination (44), we tected in nuclear extracts of IB3-1 cells after TG2 inhibition investigated whether increased TG2 protein levels might interfere (Fig. 7E). TG2 specific inhibitors such as cystamine (38) (Fig. with PPAR␥ SUMOylation. We demonstrated that blocking TG2 7, F and G), R283, or KCC009 (22) (data not shown) showed through specific gene silencing (22) increased SUMO-1 immuno- the same effects as TG2 gene silencing. Furthermore, TG2 over- reactivity in PPAR␥ immunoprecipitates in response to rosiglita- expression in A549 cells was effective in reducing PPAR␥ and zone (Fig. 7A), enhanced N-CoR protein and its nuclear localiza- IkB␣ SUMOylation (Fig. 7, H and I). TG2 SUMOylation may tion (Fig. 7B), and favored N-CoR-PPAR␥ interaction (data not therefore switch off the intracellular regulatory machinery, pre- shown). TG2 might also mediate NF-␬B activation (45) by fa- venting PPAR␥ SUMOylation, IkB␣ stabilization, and leading voring cross-linking, ubiquitination, and proteasome degrada- to an uncontrolled inflammatory response. tion of IkB␣, a key NF-␬B modulator (45) and known TG2 substrate (45). Conversely, SUMO enhancers induce SUMOy- TG2 inhibition controls inflammation in F508del-CFTR lation and stabilization of IkB␣, thus preventing NF-␬B acti- homozygous mice vation (34). We demonstrated that in IB3-1 cells the inhibition To test the effects of TG2 inhibition in vivo, we treated CF mutant of TG2 by gene silencing increased SUMO-1 immunoreactivity mice homozygous for F508del-CFTR mutation (36) and their con- in IkB␣ immunoprecipitates (Fig. 7C) and increased IkB␣ pro- trol littermates (36) with cystamine, previously reported to inhibit tein levels (Fig. 7D). Moreover, reduced p-65 NF-␬B was de- TG2 and ameliorate disease manifestations in a mouse model of 8 TG2 LINKS OXIDATIVE STRESS AND INFLAMMATION

FIGURE 9. Schematic representation of the main ﬁnding of the paper. a, Increased intracellular levels of ROS favor PIASy-TG2 interaction. b, PIASy-TG2 interaction leads to TG2 sumoylation and inhibits TG2 ubiquitination and proteasome degradation, thus leading to uncontrolled TG2 activation. c, TG2 favors PPAR␥ cross-linking and proteasome degradation and inhibits PPAR␥ SUMOylation and PPAR␥-NCoR complex interaction, thus inhibiting transrepressing activity. d, TG2 induces IkB␣ cross- linking and proteasome degradation, favoring NF-␬B activation and nuclear translocation.

Huntington’s disease (38). We treated CF and wild-type mice with SUMOylation has been defined as a key player of the posttrans- a daily injection of cystamine (i.p. injection of 100 ␮l of 0.01 M lational network to regulate key cellular functions, including tran- in PBS for 7 days) or PBS (i.p. injection of 100 ␮l of PBS for 7 scription, nuclear translocation, stress response, and chromatin days). After treatment with PBS the expression and distribution of structure, as well as of diversifying localization and even stability the tested markers remained unaltered as compared with the pat- of the modified proteins (31, 32, 46). SUMOylation is accom- tern observed in untreated CF mice. Before treatment, as well as plished via an enzymatic cascade involving, among others, E3 li- after treatment with PBS, all seven tested CF mice showed in- gases, that catalyze the transfer of SUMO from the conjugating crease of ROS (Fig. 8A) and SUMO-TG2 interaction (Fig. 8B), as enzyme UBC9 to a substrate (31). E3 ligases have gained a central well as TG2 activity (Fig. 8, C and D; D is a high magnification of role in the SUMO machinery, since they regulate SUMOylation in ␥ C), as compared with control littermates. PPAR was also reduced response to different stresses (31). Herein we demonstrate that in C D and sequestered in aggresomes (Fig. 8, and ), whereas phos- CF airway epithelia the increased levels of ROS lead to TG2 phorylation of p42–44 (Fig. 8C) and increase of TNF-␣ protein SUMOylation via interaction of TG2 with PIASy, an E3 ligase were observed (Fig. 8E). In all CF mice, the treatment with cys- already reported to mediate NEMO SUMOylation upon genotoxic tamine controlled TG2 activity (Fig. 8, C and D), increased stress through ROS generation (33). Oxidative stress increases PPAR␥ levels and its nuclear localization (Fig. 8, C and D), and PIASy protein levels and favors TG2 SUMOylation that leads to reduced p42–44 phosphorylation (Fig. 8C) and TNF-␣ protein levels (Fig. 8E), thus restoring the pattern observed in their control the persistence of high TG2 tissue levels by down-regulating TG2 littermates. Cystamine did not induce any changes in wild-type ubiquitination and proteasome degradation. mice (data not shown). Our results also indicate that TG2 SUMOylation is a general response to oxidative stress inducers since the impairment of the Discussion intracellular ROS balance, either by perturbing mitochondrial ox- In this report we have underpinned the relationship between CFTR idative regulation by rotenone or by interfering with gluthatione defective function, oxidative stress, and chronic airway inflamma- pathway, is also effective in inducing TG2 SUMOylation and en- tion in CF. We have identified TG2 SUMOylation as a pivot in hancing TG2 activity in airway epithelia. The persistence of high driving the proinflammatory CF phenotype. levels of TG2 might in turn increase the intracellular ROS since The cellular response to stress involves a finely tuned posttrans- TG2 may also contribute to the functionality of the mitochondrial lational network that provides proteins with functional ability at respiratory chain (27). Moreover, the PPAR␥ down-regulation the right time and place, and its perturbations have been shown to might also interfere with the appropriate control of the redox ma- contribute to the etiology of various human diseases (46). chinery since it can also modulate oxidative stress (47). Thus, TG2 The Journal of Immunology 9 may sustain a vicious cycle that leads to a progressive and uncon- Acknowledgments trolled impairment of the cellular homeostasis. We thank Michael Bownlee (Albert Einstein College of Medicine, New TG2 SUMOylation may therefore switch off the posttransla- York, NY) for the gift of the adenoviral vectors, G. M. Fimia (National tional regulatory mechanisms in response to the oxidative stress. Institute for Infectious Diseases, Istituto di Ricovero e Cura a Carattere Most proteins involved in the pathogenesis of chronic human dis- Scientifico, “L. Spallanzani”, Rome, Italy) for the gift of the TG2 plasmid, eases, as huntingtin, ataxin-1, tau, and ␣-synuclein, were reported Fabio Formiggini (Dynamic Imaging Microscopy, Centro di Ingegneria Genetica, Naples, Italy) for the technical support in FRET analysis, and to be SUMO (48) as well as TG2 substrates (49). PPAR␥, which Delphine Labrousse (Cancer Research UK Oncology Unit, University may be targeted by TG2 to cross-linking and proteasome deg- of Southampton, Southampton, U.K.) for the technical support in per- radation (22), may also be targeted by SUMO-1 and undergo forming experiments on mice. Cftrtm1EUR (F508del (FVB/129) mice SUMOylation in response to a PPAR␥ agonists, such as rosiglita- were obtained from Bob Scholte (Erasmus Medical Center Rotterdam, zone (22, 44). SUMOylated PPAR␥ interacts with the N-CoR- Rotterdam, The Netherlands) (European Economic Community Euro- histone deacetylase 3 (HDAC3) complex and thereby blocks its pean Coordination Action for Research in Cystic Fibrosis program EU ubiquitination, thus maintaining a repressor condition (44). Herein FP6 LSHM-CT-2005-018932). we demonstrated that sustained TG2 activation inhibits PPAR␥ SUMOylation and its interaction with the N-CoR (44), thus favor- Disclosures ing inflammation. Moreover TG2-mediated cross-linking and deg- The authors have no financial conflicts of interest. radation of IkB␣, a known TG2 substrate (45), inhibits IkB␣ SUMOylation and favors NF-␬B activation. Therefore, TG2 may References function as a link between oxidative stress and inflammation by 1. Ratjen, F., and G. Doring. 2003. Cystic fibrosis. Lancet 361: 681–689. 2. Thelin, W. R., Y. Chen, M. Gentzsch, S. M. Kreda, J. L. Sallee, C. O. Scarlett, driving the decision as to whether a protein should undergo C. H. Borchers, K. Jacobson, M. J. Stutts, and S. L. Milgram. 2007. Direct SUMO-mediated regulation or degradation (Fig. 9). interaction with filamins modulates the stability and plasma membrane expression of CFTR. J. Clin. Invest. 117: 364–374. The rheostat role of TG2 makes this enzyme an attractive 3. Gelman, M. S., and R. R. Kopito. 2003. Cystic fibrosis: premature degradation of target to restore cellular homeostasis and dampen chronic in- mutant proteins as a molecular disease mechanism. Methods Mol. Biol. 232: flammation in CF airways. The regulation of the high levels of 27–37. 4. Kopito, R. R. 1999. Biosynthesis and degradation of CFTR. Physiol. Rev. 79: TG2 protein or the inhibition of sustained TG2 enzyme activa- S167–S173. tion may represent a new attractive approach to control disease 5. Reddy, M. M., P. M. Quinton, C. Haws, J. J. Wine, R. Grygorczyk, J. A. Tabcharani, J. W. Hanrahan, K. L. Gunderson, and R. R. Kopito. 1996. evolution in CF patients. Failure of the cystic fibrosis transmembrane conductance regulator to conduct To provide the rationale and the proof-of-principle for the pu- ATP. Science 271: 1876–1879. 6. Ward, C. L., S. Omura, and R. R. Kopito. 1995. Degradation of CFTR by the tative use of TG2 inhibitors in CF patients, we checked whether ubiquitin-proteasome pathway. Cell 83: 121–127. the mechanisms described in cell lines also take place in human CF 7. Proesmans, M., F. Vermeulen, and K. De Boeck. 2008. What’s new in cystic airways. The inflammatory response is a complex event involving fibrosis?: from treating symptoms to correction of the basic defect. Eur. J. Pe- diatr. 167: 839–849. different cell types interacting within their natural environment. 8. Verkman, A. S., and L. J. Galietta. 2009. Chloride channels as drug targets. Nat. We took advantage of a well-established model of in vitro cultures Rev. Drug Discov. 8: 153–171. 9. Caputo, A., E. Caci, L. Ferrera, N. Pedemonte, C. Barsanti, E. Sondo, U. Pfeffer, of explants of CF nasal polyps, which are routinely removed sur- R. Ravazzolo, O. Zegarra-Moran, and L. J. Galietta. 2008. TMEM16A, a mem- gically and represent CF chronic airway inflammation (22, 41). We brane protein associated with calcium-dependent chloride channel activity. Sci- demonstrated that ROS-mediated PIASy-TG2 as well as SUMO- ence 322: 590–594. 10. Scheid, P., L. Kempster, U. Griesenbach, J. C. Davies, A. Dewar, P. P. Weber, 1-TG2 interactions take place in human CF airways. W. H. Colledge, M. J. Evans, D. M. Geddes, and E. W. Alton. 2001. Inflamma- To evaluate whether we can translate our in vitro findings into tion in cystic fibrosis airways: relationship to increased bacterial adherence. Eur. Respir. J. 17: 27–35. an appropriate animal model and confirm their biological rele- 11. Smith, J. J., S. M. Travis, E. P. Greenberg, and M. J. Welsh. 1996. Cystic fibrosis vance in vivo we have studied CF mutant mice homozygous for airway epithelia fail to kill bacteria because of abnormal airway surface fluid. F508del-CFTR (36). We have demonstrated TG2-SUMO colocal- Cell 85: 229–236. 12. Kowalski, M. P., A. Dubouix-Bourandy, M. Bajmoczi, D. E. Golan, T. Zaidi, ization and increased TG2 activity in lung tissues from F508del- Y. S. Coutinho-Sledge, M. P. Gygi, S. P. Gygi, E. A. Wiemer, and G. B. Pier. CFTR homozygous mice. In the lungs of these mice, PPAR␥ is 2007. Host resistance to lung infection mediated by major vault protein in epithelial cells. Science 6: 130–132. also reduced and sequestered in aggresomes, and inflammation is 13. Thelin, W. R., and R. C. Boucher. 2007. The epithelium as a target for therapy present. We treated these CF mice with daily i.p. injections of in cystic fibrosis. Curr. Opin. Pharmacol. 7: 290–295. cystamine, already used in a mouse model of Huntington’s disease 14. Escotte, S., O. Tabary, D. Dusser, C. Majer-Teboul, E. Puchelle, and J. Jacquot. 2003. Fluticasone reduces IL-6 and IL-8 production of cystic fibrosis bronchial (38). Cystamine is known to inactivate TG2 through a disulfide- epithelial cells via IKK-␤ kinase pathway. Eur. Respir. J. 21: 574–581. exchange reaction and is a substrate for TG2 (38). We have shown 15. Rottner, M., C. Kunzelmann, M. Mergey, J. M. Freyssinet, and M. C. Martinez. 2007. Exaggerated apoptosis and NF-␬B activation in pancreatic and tracheal that daily injections of cystamine inhibit TG2 activation, increase cystic fibrosis cells. FASEB J. 21: 2939–2948. PPAR␥ protein expression, and control inflammation. 16. Venkatakrishnan, A., A. A. Stecenko, G. King, T. R. Blackwell, K. Brigham, Our results highlight TG2 as an unforeseen unifying link be- J. W. Christman, and T. S. Blackwell. 2000. Exaggerated activation of nuclear factor-␬B and altered I␬B-beta processing in cystic fibrosis bronchial epithelial tween genetic defect, deregulation of cellular homeostasis, and in- cells. Am. J. Respir. Cell Mol. Biol. 23: 396–403. flammation (Fig. 9). They also indicate TG2 as a candidate target 17. Becker, M. N., M. S. Sauer, M. S. Muhlebach, A. J. Hirsh, Q. Wu, M. W. Verghese, and S. H. Randell. 2004. Cytokine secretion by cystic fibrosis for the design of a pathogenic-based therapy in CF and add to the airway epithelial cells. Am. J. Respir. Crit. Care Med. 169: 645–653. rationale for attempting inhibition of TG2 in CF patients. This 18. Hybiske, K., Z. Fu, C. Schwarzer, J. Tseng, J. Do, N. Huang, and T. E. Machen. 2007. Effects of cystic fibrosis transmembrane conductance regulator and study also suggests that targeting TG2 SUMOylation through ⌬F508CFTR on inflammatory response, ER stress, and Ca2ϩ of airway epithelia. inhibition of TG2-SUMO interactions might be helpful to con- Am. J. Physiol. 293: L1250–L1260. trol the unwanted persistence of TG2, thus favoring TG2 ubiq- 19. Teichgra¨ber, V., M. Ulrich, N. Endlich, J. Riethmu¨ller, B. Wilker, C. C. De Oliveira-Munding, A. M. van Heeckeren, M. L. Barr, G. von Ku¨rthy, uitination and proteasome degradation. Therefore, TG2 inhibi- K. W. Schmid, et al. 2008. Ceramide accumulation mediates inflammation, cell tion might represent a new attractive option to control the death and infection susceptibility in cystic fibrosis. Nat. Med. 14: 382–391. 20. Perez, A., A. C. Issler, C. U. Cotton, T. J. Kelley, A. S. Verkman, and P. B. Davis. evolution of chronic inflammatory diseases, neurodegenerative 2007. CFTR inhibition mimics the cystic fibrosis inflammatory profile. diseases, and even cancer. Am. J. Physiol. 292: L383–L395. 10 TG2 LINKS OXIDATIVE STRESS AND INFLAMMATION

21. Verhaeghe, C., C. Remouchamps, B. Hennuy, A. Vanderplasschen, A. Chariot, 37. Nicklas, W., P. Baneux, R. Boot, T. Decelle, A. A. Deeny, M. Fumanelli, and B. S. Tabruyn, C. Oury, and V. Bours. 2007. Role of IKK and ERK pathways in Illgen-Wilcke; for FELASA (Federation of European Laboratory Animal Science intrinsic inflammation of cystic fibrosis airways. Biochem. Pharmacol. 73: Associations Working Group on Health Monitoring of Rodent and Rabbit Col- 1982–1994. onies). 2002. Recommendation for the health monitoring of rodent and rabbit 22. Maiuri, L., A. Luciani, I. Giardino, V. Raia, V. R. Villella, M. D’Apolito, colonies in breeding and experimental units. Lab. Anim. 36: 20–42. M. Pettoello-Mantovani, S. Guido, C. Ciacci, M. Cimmino, et al. 2008. Tissue 38. Karpuj, M. V., M. W. Becher, J. E. Springer, D. Chabas, S. Youssef, R. Pedotti, transglutaminase activation modulates inflammation in cystic fibrosis via PPAR␥ D. Mitchell, and L. Steinman. 2002. Prolonged survival and decreased abnormal down-regulation. J. Immunol. 180: 7697–7705. movements in transgenic model of Huntington disease, with administration of the 23. Rottner, M., J. M. Freyssinet, and M. C. Martínez. 2009. Mechanisms of the transglutaminase inhibitor cystamine. Nat. Med. 8: 143–149. noxious inflammatory cycle in cystic fibrosis. Respir. Res. 10: 23. 39. Kenworthy, A. K. 2001. Imaging protein-protein interactions using fluorescence ␬ 24. Vij, N., S. Mazur, and P. L. Zeitlin. 2009. CFTR is a negative regulator of NF B resonance energy transfer microscopy. Methods 24: 289–296. mediated innate immune response. PLoS ONE 4: e4664. 40. Mu¨ller, S., and C. Hoege. 2001. SUMO-1, ubiquitin’s mysterious cousin. Nat. 25. Lorand, L., and R. M. Graham. 2003. Transglutaminases: crosslinking enzymes Rev. Mol. Cell Biol. 2: 202–210. with pleiotropic functions. Nat. Rev. Mol. Cell Biol. 4: 140–156. 26. Daynes, R. A., and D. C. Jones. 2002. Emerging roles of PPARs in inflammation 41. Raia, V., L. Maiuri, C. Ciacci, I. Ricciardelli, L. Vacca, S. Auricchio, and immunity. Nat. Rev. Immunol. 2: 748–759. M. Cimmino, M. Cavaliere, M. Nardone, A. Cesaro, et al. 2005. Inhibition of p38 27. Malorni, W., M. G. Farrace, C. Rodolfo, and M. Piacentini. 2008. Type 2 trans- mitogen activated protein kinase controls airway inflammation in cystic fibrosis. glutaminase in neurodegenerative diseases: the mitochondrial connection. Curr. Thorax 60: 773–780. Pharm. Des. 14: 278–288. 42. Martinvalet, D., D. M. Dykxhoorn, R. Ferrini, and J. Lieberman. 2008. Granzyme 28. Lorand, L. 1996. Neurodegenerative diseases and transglutaminase. Proc. Natl. A cleaves a mitochondrial complex I protein to initiate caspase-independent cell Acad. Sci. USA 93: 14310–14313. death. Cell 133: 681–692. 29. Molberg, O., S. N. Mcadam, R. Ko¨rner, H. Quarsten, C. Kristiansen, L. Madsen, 43. Muniyappa, H. B., S. Song, C. K. Mathews, and K. C. Das. 2009. Reactive L. Fugger, H. Scott, O. Noreń, P. Roepstorff, et al. 1998. Tissue transglutaminase oxygen species-independent oxidation of thioredoxin in hypoxia: inactivation of selectively modifies gliadin peptides that are recognized by gut-derived T cells in ribonucleotide reductase and redox-mediated checkpoint control. J. Biol. Chem. celiac disease. Nat. Med. 4: 713–717. 284: 17069–17081. 30. Verna, A., H. Wang, B. Manavathi, J. Y. Fok, A. P. Mann, R. Kumar, and 44. Pascual, G., A. L. Fong, S. Ogawa, A. Gamliel, A. C. Li, V. Perissi, D. W. Rose, K. Mehta. 2006. Increased expression of tissue transglutaminase in pancreatic T. M. Willson, M. G. Rosenfeld, and C. K. Glass. 2005. A SUMOylation-de- ductal adenocarcinoma and its implications in drug resistance and metastasis. pendent pathway mediates transrepression of inflammatory response genes by Cancer Res. 66: 10525–10533. PPAR-␥. Nature 437: 759–763. 31. Meulmeester, E., and F. Melchior. 2008. Cell biology: SUMO. Nature 452: 45. Kim, D. S., S. S. Park, B. H. Nam, I. H. Kim, and S. Y. Kim. 2006. Reversal of 709–711. drug resistance in breast cancer cells by transglutaminase 2 inhibition and nuclear 32. Geiss-Friedlander, R., and F. Melchior. 2007. Concepts in sumoylation: a decade factor-␬B inactivation. Cancer Res. 66: 10936–10943. on. Nat. Rev. Mol. Cell Biol. 8: 947–956. 46. Tempe`, D., M. Piechaczyk, and G. Bossis. 2008. SUMO under stress. Biochem. 33. Mabb, A. M., and S. M. Wuerzberger-Davis. 2006. PIASy mediates NEMO Soc. Trans. 36: 874–878. sumoylation and NF-␬B activation in response to genotoxic stress. Nat. Cell Biol. 8: 986–993. 47. Quintanilla, R. A., Y. N. Jin, K. Fuenzalida, M. Bronfman, and G. V. Johnson. 34. Carbia-Nagashima, A., J. Gerez, C. Perez-Castro, M. Paez-Pereda, S. Silberstein, 2008. Rosiglitazone treatment prevents mitochondrial dysfunction in mutant hun- G. K. Stalla, F. Holsboer, and E. Arzt. 2007. RSUME, a small RWD-containing tingtin-expressing cells: possible role of peroxisome proliferator-activated recep- ␥ ␥ protein, enhances SUMO conjugation and stabilizes HIF-1␣ during hypoxia. Cell tor- (PPAR ) in the pathogenesis of Huntington disease. J. Biol. Chem. 283: 131: 309–323. 25628–25637. 35. Du, X., D. Edelstein, S. Obici, N. Higham, M. H. Zou, and M. Brownlee. 2006. 48. Steffan, J. S., N. Agrawal, J. Pallos, E. Rockabrand, L. C. Trotman, N. Slepko, Insulin resistance reduces arterial prostacyclin synthase and eNOS activities by K. Illes, T. Lukacsovich, Y. Z. Zhu, E. Cattaneo, et al. 2004. SUMO modification increasing endothelial fatty acid oxidation. J. Clin. Invest. 116: 1071–1079. of huntingtin and Huntington’s disease pathology. Science 304: 100–104. 36. Legssyer, R., F. Huaux, J. Lebacq, M. Delos, E. Marbaix, P. Lebecque, D. Lison, 49. Juun, E., R. D. Ronchetti, M. M. Quezado, S. Y. Kim, and M. M. Mouradian. B. J. Scholte, P. Wallemacq, and T. Leal. 2006. Azithromycin reduces sponta- 2003. Tissue transglutaminase-induced aggregation of ␣-synuclein: implication neous and induced inflammation in ⌬F508 cystic fibrosis mice. Respir. Res. for Lewy body formation in Parkinson’s disease and dementia with Lewy bodies. 7: 134. Proc. Natl. Acad. Sci. USA 100: 2047–2052. The Journal of Immunology

Tissue Transglutaminase Activation Modulates Inﬂammation in Cystic Fibrosis via PPAR␥ Down-Regulation1

Luigi Maiuri,2*† Alessandro Luciani,†‡ Ida Giardino,§ Valeria Raia,¶ Valeria R. Villella,ʈ Maria D’Apolito,† Massimo Pettoello-Mantovani,† Stefano Guido,ʈ Carolina Ciacci,‡ Mariano Cimmino,# Olivier N. Cexus,** Marco Londei,††‡‡ and Sonia Quaratino2**††

Cystic fibrosis (CF), the most common life-threatening inherited disease in Caucasians, is due to mutations in the CF transmembrane conductance regulator (CFTR) gene and is characterized by airways chronic inflammation and pulmonary infections. The inflammatory response is not secondary to the pulmonary infections. Indeed, several studies have shown an increased proinflammatory activity in the CF tissues, regardless of bacterial infections, because inflammation is similarly observed in CFTR-defective cell lines kept in sterile conditions. Despite recent studies that have indicated that CF airway epithelial cells can spontaneously initiate the inflammatory cascade, we still do not have a clear insight of the molecular mechanisms involved in this increased inflammatory response. In this study, to understand these mechanisms, we investigated ex vivo cultures of nasal polyp mucosal explants of CF patients and controls, CFTR-defective IB3-1 bronchial epithelial cells, C38 isogenic CFTR corrected, and 16HBE normal bronchial epithelial cell lines. We have shown that a defective CFTR induces a remarkable up-regulation of tissue transglutaminase (TG2) in both tissues and cell lines. The increased TG2 activity leads to functional sequestration of the anti-inflammatory peroxisome proliferator-activated receptor ␥ and increase of the classic parameters of inflammation, such as TNF-␣, tyrosine phosphorylation, and MAPKs. Specific inhibition of TG2 was able to reinstate normal levels of peroxisome proliferator-activated receptor-␥ and dampen down inflammation both in CF tissues and CFTR-defective cells. Our results highlight an unpredicted central role of TG2 in the mechanistic pathway of CF inflammation, also opening a possible new wave of therapies for sufferers of chronic inflammatory diseases. The Journal of Immunology, 2008, 180: 7697–7705.

he breach of the innate immune system and a chronically The direct link between CF transmembrane conductance regu- infected and inflamed lung are the characteristic features lator (CFTR) dysfunction and airway infection has been shown. T of cystic fibrosis (CF)3 (1). It has been estimated that Defects in the CFTR result in an impaired clearance of lung patho- ϳ30,000 children and adults in the U.S. (70,000 worldwide) are gens, especially Pseudomonas aeruginosa (2, 3). Infections are affected by CF, with a frequency of ϳ1 in 2,500 livebirths (1). then further exacerbated by the poor mucociliary clearance due to ciliary dysfunction and the hyperabsorption of water by the airways epithelium (4). *European Institute for Cystic Fibrosis Research, San Raffaele Scientific Institute, The mechanisms by which CFTR mutations might contribute to Milan, Italy; †Institute of Pediatrics, University of Foggia, Foggia, Italy; ‡Department airways inflammation are, however, still elusive. Defects of the § of Experimental Medicine, University Federico II, Naples, Italy; Department of CFTR are also associated with a marked increase of proinflamma- Laboratory Medicine, University of Foggia, Foggia, Italy; ¶Department of Pedi- atrics, University Federico II, Naples, Italy; ʈDepartment of Chemical Engineer- tory cytokines, such as TNF-␣, IL-6, IL-1␤, IL-17 (5, 6), and the ing, University Federico II, Naples, Italy; #Department of Otolaryngology, Uni- potent neutrophil chemoattractant and activator IL-8, which re- versity Federico II, Naples, Italy; **Cancer Research UK Oncology Unit, University of Southampton, Southampton, United Kingdom; ††Institute of Child cruits large numbers of neutrophils into the airways (1). Chronic Health, University College, London, United Kingdom; and ‡‡Novartis Pharma AG inflammation, however, is not merely the consequence of repeti- Translational Medicine, Basel, Switzerland tive infections, because CFTR-defective cell lines spontaneously Received for publication January 8, 2008. Accepted for publication March develop similar proinflammatory features when kept in vitro in 28, 2008. sterile conditions (7, 8). Indeed, CFTR-defective cells have been The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance reported to have an intrinsically proinflammatory phenotype, and with 18 U.S.C. Section 1734 solely to indicate this fact. despite different intracellular pathways that have been considered, 1 This work was supported by the European Institute for Cystic Fibrosis Research, none to date has been demonstrated to be associated (7, 8). Cancer Research U.K., and Rothschild Trust. To underpin the molecular link between a defected CFTR and 2 Address correspondence and reprint requests to Dr. Sonia Quaratino, Cancer Re- the excessive inflammatory responses typical of CF airways here search UK Oncology Unit, University of Southampton, SO16 6YD Southampton, we have studied nasal polyp mucosa from CF patients (9) as well UK. E-mail address: [email protected] or Luigi Maiuri, European Institute for Cystic Fibrosis Research, San Raffaele Scientific Institute, via Olgettina 58, 20132, Milan, as CFTR-defective bronchial epithelial cell lines (2, 10). The ra- Italy. E-mail address: [email protected] tionale to extend the study to the cell lines was to specifically 3 Abbreviations used in this paper: CF, cystic fibrosis; bio-MDC, biotinylated mono- target the role of the defective CFTR on peroxisome proliferator- dansylcadaverine; CFTR, CF transmembrane conductance regulator; CM- ␥ ␥ H2DCFDA, 5(and 6)-chloromethyl-2Ј7Ј-dichlorodihydrofluorescein diacetate acetyl activated receptor- (PPAR ) in sterile conditions, excluding the ester; HDAC6, histone deacetylase 6; NAC, N-acetyl-cysteine; PPAR␥, peroxisome influence that recurrent infections might have on the epithelium proliferator-activated receptor-␥; ROS, reactive oxygen species; siRNA, small inter- and the generation of inflammation. fering RNA; TG2, tissue transglutaminase. We have highlighted how a defected CFTR induces high levels Copyright © 2008 by The American Association of Immunologists, Inc. 0022-1767/08/$2.00 of tissue transglutaminase (TG2) in the airways. TG2 has a pivotal www.jimmunol.org 7698 TG2 DRIVES INFLAMMATION VIA PPAR␥ DOWN-REGULATION role in the initiation of inflammation by sequestering PPAR␥,a with Wallac 1420 multilabel counter (PerkinElmer). Tissue sections (5 nuclear hormone receptor expressed in monocytes, macrophages, ␮m) of nasal mucosa and cells were incubated with 10 ␮M CM- and epithelial cells that negatively regulates inflammatory gene H2DCFDA, according to manufacturer suggestions, and an LSM510 Zeiss confocal laser-scanning unit (Carl Zeiss, Germany) was used for detection. expression by transrepressing inflammatory responses (11). This work also suggests that TG2 inhibition could become a therapeutic In situ detection of TG2 activity and protein target to control inflammation in CF and possibly in other chronic TG2 protein and enzyme activity were detected and analyzed by confocal inflammatory diseases. microscopy. Briefly, for detection of TG2 enzyme activity, live cells were preincubated with TG2 assay buffer (965 ␮l of 100 mM Tris/HCl (pH 7.4), 25 ␮l of 200 mM CaCl ) for 15 min and then with the same TG2 assay Materials and Methods 2 buffer added with 10 ␮l of 10 mM biotinylated monodansylcadaverine Human airway biopsies and ex vivo cultures (bio-MDC; Molecular Probes) for1hatroom temperature. The reaction Nasal polyp explants from 10 CF patients carrying the common CFTR mu- was stopped with 25 mM EDTA for 5 min. The cells were then fixed in 4% tations (⌬F508/⌬F508, ⌬F508/W1282X, ⌬F508/N1303K, or ⌬F508/G542X) paraformaldehyde for 10 min. The incorporation of labeled substrate and 10 non-CF patients with nonallergic idiopathic polyposis were cultured, was visualized by incubation with PE-conjugated streptavidin (Dako- for 4–24 h (9), with or without specific TG2 inhibitors 1,3-dymethyl-2-[(2- Cytomation; 1:50) for 30 min. Control experiments included the omis- oxopropyl) thio] imidazolium (R283) (250 ␮M) (12) or halo-dihydroisox- sion of bio-MDC, as well as replacement of 200 mM CaCl2 with 200 ␮ mM EDTA. Blocking experiments were also conducted by incubating azole-derivate transglutaminase inhibitor KCC009 (250 M), reactive oxygen ␮ species (ROS) scavenger EUK 134 (50 ␮g/ml; Alexis Biochemical), N-acetyl- cells with the active site inhibitor R283 (250 M) for 1 h before incu- cysteine (NAC, 10 mM; Alexis Biochemical), PPAR␥ antagonist GW9662 (1 bation with the substrates (bio-MDC or biotinylated peptides) and de- ␮M; Alexis Biochemical), or R283 for 24 h, followed by GW9662 (1 ␮M) for tection of TG2 enzymatic activity. Data were analyzed under fluores- 4 h. Informed consent was obtained from all subjects, and the ethical com- cence examination by LSM510 Zeiss confocal laser-scanning unit (Carl mittee of Regione Campania Health Authority approved the study. Zeiss). The simultaneous detection of TG2 protein and activity was performed Cell lines and cultures by incubating live cells with bio-MDC for1hatroom temperature, according to the above described procedure, and then with anti-TG2 CUB IB3-1 (human CF bronchial epithelial cell line with the common ⌬F508/ 7402 mAb (NeoMarkers; 1:50) for1hatroom temperature. This was W1282X CFTR mutation) and C38 (isogenic stably rescued with functional followed by simultaneous incubation with PE-conjugated streptavidin Ј CFTR) cell lines (LGC Promochem) (2, 7, 10) were stimulated for 6 h with (DakoCytomation; 1:50) and FITC-conjugated rabbit anti-mouse Ig F(ab )2 R283 (250 ␮M) or KCC009 (250 ␮M), ionomycin (1 ␮M; Calbiochem), (DakoCytomation; 1:20) for 1 h. BAPTA-AM (5 ␮M, Calbiochem), EUK 134 (50 ␮g/ml), rosiglitazone (10 The detection of TG2 activity and protein on tissue sections was per- ␮␮), NAC (10 mM), proteasome inhibitor MG132 (50 ␮M for 6 h; Calbio- formed, as reported in our previous paper (12). chem), or R283 for 24 h, followed by 6-h rosiglitazone. Normal human bronchial epithelial 16HBE cells were cultured, as previously described (8). Confocal microscopy Quantitative RT-PCR Cell lines were fixed in methanol and permeabilized with 0.5% Triton X-100 prior incubation with primary Abs. Frozen tissue sections were fixed Quantitative RT-PCR was performed using iCycler iQ Multicolour Real- in acetone for 10 min. Anti-phospho-p42/p44 (1:500; Cell Signaling Tech- Time PCR Detector (Bio-Rad) with iQ TM SYBR Green supermix (Bio- nology), anti-phosphotyrosine PY99 mAb (1:80, mouse IgG2b), anti- Rad). A relative quantitative method was applied for TG2 mRNA (Qiagen; PPAR␥ (clone E8, sc-7273, 1:100, mouse IgG1; clone H100, sc-7196, catalog QT00081277), normalized by the control GAPDH mRNA. 1:100, rabbit polyclonal IgG; Santa Cruz Biotechnology), anti-ubiquitin (1:100 clone FL-76, rabbit polyclonal IgG), anti-histone deacetylase 6 RNA interference (HDAC6; 1:100 clone H300, rabbit polyclonal IgG), anti-p65 NF-␬B ␥ (F-6; 1:100), and anti-ICAM-1 (15.2; 1:200) (Santa Cruz Biotechnol- IB3-1 cells were transfected with 50 nM human TG2, human PPAR , and ogy) Abs were used for first Ab. The Ag expression and distribution scramble small interfering RNAs (siRNAs) duplex using Hiperfect Trans- were visualized by indirect immunofluorescence, as previously de- fection Reagent (Qiagen) at 37°C for 72 h. The target sequence of TG2 scribed (12). siRNA was 5Ј-CCGCGTCGTGACCAACTACAA-3Ј, and the whole-cell lysate was then analyzed for Western blot. The PPAR␥ siRNA is a pool of ELISA three sequences, as follows: CCAAGUAACUCUCCUCAAAtt, GAAUG UGAAGCCCAUUGAAtt, and CUACUGCAGGUGAUCAAGAtt. Trans- Human TNF-␣ secretion was measured using the BD OptEIATM TNF-␣ fected cells were then analyzed by confocal microscopy for expression ELISA kit II (BD Biosciences). Measurements were performed at least in 6 using anti-PPAR␥ mAb clone E8. triplicate. Values were normalized to 10 cells; results were expressed as mean Ϯ SEMs. Western blot Statistical analysis First Abs anti-phospho-p42/p44 MAPKs (Cell Signaling Technology), PPAR␥ (clone E8 sc-7273; Santa Cruz Biotechnology), TG2 (clone All experiments were performed at least in triplicate. Data distribution was ␧ ␥ analyzed, and statistical differences were evaluated by using ANOVA CUB7402; DakoCytomation), N ( -L-glutamyl)-L-lysine isopeptide Ͻ (clone 81DIC2; Covalab), and ubiquitin (1:100 clone FL-76, rabbit Tukey-Kramer test by SPSS 12 software. A p value of 0.05 was consid- polyclonal IgG) were counterstained by a HRP-conjugated anti-IgG Ab ered significant. (Amersham, General Healthcare). The amounts of proteins were determined by a Bio-Rad protein assay to ensure equal protein loading be- Results fore Western blot analysis. Fifty micrograms of cell lysate were loaded PPAR␥ colocalizes in perinuclear aggresomes with ubiquitin in each lane. and HDAC6 in CF epithelium Immunoprecipitation We initially determined whether PPAR␥, a key inflammation reg- Following cell lysis, 500 ␮g of proteins was incubated at 4°C overnight ulator implicated in the prevention of several immunoinflamma- with anti-PPAR␥ Ab (clone E8) and then mixed with protein G-Sepha- tory disorders (13), had any role in CF. PPAR␥, produced by sev- rose beads for 2 h. The beads were washed three times with lysis buffer, eral cell types, including epithelial cells, regulates inflammation and immunoprecipitates were resuspended in SDS loading buffer. Equal via the specific inhibition of proinflammatory cytokines, such as amounts of immunoprecipitate were loaded and analyzed for TNF-␣, IL-6, IL-1␤, and the NF-␬B pathway, and by modulating Western blot. oxidative stress (13–15). ROS detection Nasal polyp mucosal biopsies from CF patients and controls Cell lines were pulsed with 10 ␮M 5(and 6)-chloromethyl-2Ј7Ј-dichlorodi- were ex vivo cultured, as previously reported, so that epithelial, hydrofluorescein diacetate acetyl ester (CM-H2DCFDA; Molecular myeloid, and lymphoid components can retain all the interactions Probes, Invitrogen), according to manufacturer suggestions, and analyzed with neighboring cells within their natural environment (9). The Journal of Immunology 7699

FIGURE 1. PPAR␥ expression in airway epithelial cells. Confocal images of nasal polyp mucosa from CF patients and controls (A) and IB3-1 and C38 cells (B): low PPAR␥ expression (green) (as detected by clone E8 Ab) with perinuclear aggregates in nasal CF epithelia and in IB3-1 cells; CyTRAK Orange (red), nuclear counterstaining. Yellow color indicates nuclear localization of PPAR␥. C, Confocal images of IB3-1 cells: co- localization (yellow) of ubiquitin (red, top line)or HDAC6 (red, bottom line) in perinuclear aggregates of PPAR␥ (green). D, Left panel, Immunoblot detection of PPAR␥ in C38 and IB3-1 cells incubated with or without the proteasome inhibitor MG132. Right panel, Im- munoprecipitated PPAR␥ species from whole-cell extracts of IB3-1 cells are immunoreactive for the anti- ubiquitin Ab. Immunoprecipitation (IP): anti-PPAR␥ Ab; immunoblot (IB): anti-ubiquitin Ab. E, Confocal images of IB3-1 cells: high PPAR␥ expression (green) in perinuclear aggregates upon MG132 incubation. F, IB3-1 cells after incubation with MG132: expression and colocalization of ubiquitin/PPAR␥ (yellow, left panel) or HDAC6/PPAR␥ (yellow, right panel)inag- gresomes. A–C and E and F, Scale bar, 10 ␮m.

We found that PPAR␥ expression in nasal biopsies from CF CFTR-defective IB3-1 cell line, the perinuclear aggregates of patients and the CFTR-defective IB3-1 cell line was significantly PPAR␥ also colocalize with both HDAC6 and ubiquitin (Fig. 1C), reduced compared with controls and the isogenic C38 cell line thus demonstrating the aggresome nature of these aggregates. Be- with functional CFTR (Fig. 1, A and B). Remarkably, PPAR␥ was cause the presence of aggresomes and HDAC6 is usually the cel- mainly limited to perinuclear localization in CF tissues and CFTR- lular response to eliminate cytoplasmic misfolded proteins (17), defective IB3-1 cell line (Fig. 1, A and B). we have then investigated whether PPAR␥ was subjected to pro- This perinuclear localization was suggestive of aggresomes, a teasome degradation. Treatment with the specific proteasome in- specific cellular response against accumulation of ubiquitinated hibitor MG132 (18) led to an increase in PPAR␥ protein levels misfolded and/or aggregated proteins (16), which might also result (Fig. 1, D and E) and PPAR␥ ubiquitination (Fig. 1D) in CFTR- from impaired proteasome function (16). Although aggresomes defective IB3-1 cells. Treatment with MG132 also led to an im- containing PPAR␥ have not been described before, aggresomes pressive perinuclear increase of HDAC6- and ubiquitin-PPAR␥ containing ubiquitinated misfolded ⌬F508 CFTR molecules have colocalization (Fig. 1F) in IB3-1 cells, suggesting a role for pro- been previously observed as result of overexpression of the defec- teasome degradation. tive CFTR (16). The aggresomes of CFTR also contain HDAC6, a Similar results were obtained using a polyclonal anti-PPAR␥ Ab microtubule-associated deacetylase that interacts with ubiquitin with different epitope specificity (data not shown). To confirm the and stabilizes polyubiquitin chains (17). We have shown that in the anti-PPAR␥ Ab specificity, IB3-1 cells were transfected with 7700 TG2 DRIVES INFLAMMATION VIA PPAR␥ DOWN-REGULATION

(20–23). TG2 is a pleiotropic enzyme expressed by many cell types, including epithelial cells, often up-regulated in many chronic inflammatory conditions (22, 23). Therefore, we explored whether TG2 levels or its enzymatic activity were up-regulated in CF tissues and CFTR-defective cell lines. We observed a significant increase of TG2 protein and enzymatic activity in CF tissues as well as the IB3-1 cell line (Fig. 2, A–C). The increase of TG2 was demonstrated by immunoblot and confocal microscopy, whereas real-time PCR showed that TG2 transcript was significantly higher in IB3-1 than C38 cell lines (Fig. 2D). These results clearly indicate that CFTR-defective epithelial cells are characterized by an intrinsic TG2 functional increase.

TG2 induces cross-linking of PPAR␥ and mediates aggresome formation We then tested the hypothesis that TG2 might induce posttranslational modifications (cross-linking) of PPAR␥ in CFTR-defective cells, because the QG and QXXP motifs in the PPAR␥ sequence could be recognized as specific sites for TG2 activity (22). Double labeling immunofluorescence with anti-PPAR␥ Ab and an isopeptide cross-link specifically catalyzed by TG2 (22) demonstrated colocalization only in IB3-1 perinuclear aggregates (Fig. 3A). Fur- thermore, using an anti-PPAR␥ Ab, several high molecular mass bands ranging between 72 and 250 kDa were immunoprecipitated and detected by the anti-isopeptide Ab in the IB3-1 line (Fig. 3B). Normal 55-kDa PPAR␥ in IB3-1 cells was reduced (Fig. 3C), whereas C38 cell line presented only faint levels of the high molecular mass and a considerable amount of the normal 55-kDa PPAR␥ (Fig. 3, B and C). Exposure of the IB3-1 cells to the irreversible TG2-specific inhibitor R283 (a gift from M. Griffin, Aston University, Birmingham, U.K.) (12, 23, 24) tested at increasing concentrations 50–500 ␮M with similar results and induced a significant reduction of high molecular mass PPAR␥ (Fig. 3D) and a significant increase of the normal 55-kDa PPAR␥ in the IB3-1 cells (Fig. 3E). This indicated that the crosslinked PPAR␥ was due to increased TG2 activity. A similar result was achieved through TG2 gene silencing by siRNA (Fig. 3F). Exposure of both CF tissues and IB3-1 cells to R283 also resulted in a significant reduction of PPAR␥ aggregates, restoring a wider intracellular distribution of PPAR␥ (Fig. 3G). In IB3-1 cells, PPAR␥ nuclear translocation was promoted only by R283, whereas rosiglitazone, a PPAR␥ agonist with potent anti- inflammatory activity used in clinic (25), induced only a marginal nuclear PPAR␥ increase (Fig. 3H). In contrast, rosiglitazone alone was able to induce an impressive PPAR␥ nuclear translocation in C38 cells (Fig. 3H). FIGURE 2. TG2 expression in airway epithelial cells. Confocal im- These results indicate that PPAR␥ aggregates in CFTR-defec- ages identified expression of TG2 (green) and its activity (blue) in epithelial cells of nasal polyp mucosa from CF patients (A) and IB3-1 CF tive cells are induced by TG2. cells (B) as compared with controls. Merge indicates overlay of individual channels (cyan). Scale bar, 10 ␮m. C, Lysates of C38 and IB3-1 TG2 inhibition controls inflammation in CF epithelial cells cells were immunoblotted with anti-TG2 Abs. D, Real-time PCR of TG2 activity does not only control the levels and function of PPAR␥, TG2 mRNA in C38 and IB3-1 cells. TG2 transcript was significantly p Ͻ 0.05 vs C38 cells). Error bars but also influences other markers of inflammation. Inhibition of TG2 ,ء) higher in IB3-1 than C38 cells represent SEMs. with either R283 or KCC009 (another TG2-specific inhibitor, gift from C. Khosla, Stanford University, Palo Alto, CA; data not shown) induced a significant reduction of the phosphorylated p42/p44 human PPAR␥ siRNA and then immunostained with anti-PPAR␥ MAPKs (Fig. 4A), ICAM-1, and NF-␬B activation (data not shown) mAb clone E8. We did not detect any PPAR␥ in the IB3-1 cell both in CF tissues and IB3-1 cells. Inhibition of TG2 by siRNA also transfected with human PPAR␥ siRNA (data not shown). induced a significant decrease of the phosphorylated p42/p44 (Fig. 4B) and TNF-␣ release (Fig. 4C) in IB3-1 cells. TG2 is up-regulated in human CFTR-defective cells The inhibitory activity of R283 on tyrosine phosphorylation, a Aggresomes are a prominent cytopathological feature of most neu- well-established marker of epithelial inflammation in human CF rodegenerative disorders, including Parkinson’s and Huntington’s airway mucosa (9), was abrogated by 4-h incubation with the diseases (19). In these pathologies, there is also an increase in PPAR␥ antagonist GW9662 (Fig. 4D), further indicating that TG2 TG2, responsible for the aggregation of ␣-synuclein in Parkinson’s controls inflammation through PPAR␥. The Journal of Immunology 7701

FIGURE 3. Effects of TG2 on PPAR␥ protein expression and localization. A, Confocal images of IB3-1 and C38 cells immunostained with PPAR␥ (green) and the anti- isopeptide Ab (red). Merge of individual channels shows colocalization (yellow) in perinuclear aggregates. B, Immunoprecipitated PPAR␥ species from whole-cell extracts of C38 and IB3-1 cells are immunoreactive for the anti-isopeptide cross-link Ab. Im- munoprecipitation (IP): anti-PPAR␥ Ab; immunoblot (IB): anti-isopeptide Ab. C, Immunoblot of PPAR␥ in C38 and IB3-1 cells. D, Effect of R283 on PPAR␥ immunoreactivity for the anti- isopeptide Ab in IB3-1 cells. IP: anti- PPAR␥ Ab; IB: anti-isopeptide Ab. E, Immunoblot analysis of PPAR␥ in IB3-1 cells with or without R283. F, Immunoblot of PPAR␥ in IB3-1 cells: effect of TG2 siRNA. Scramble siRNA was used as a negative control. 72 kDa, TG2; 55 kDa, PPAR␥. Con- focal images of G, PPAR␥ (green) immunostaining of CF nasal mucosa and IB3-1 cells with or without R283, and H, PPAR␥ (green) in IB3-1 cells (left panel) and in C38 cells (right panel) after incubation with PPAR␥ agonist rosiglitazone (10 ␮Mfor6h) with or without prior incubation with R283. CyTRAK Orange (red) nuclear counterstaining. Yellow color indicates nuclear localization of PPAR␥. A, G, and H, Scale bar, 10 ␮m.

Intracellular Ca2ϩ and ROS levels modulate TG2 activity and Ca2ϩ decrease, induced by the Ca2ϩ chelator BAPTA-AM inflammation in CFTR-defective cells (27), significantly reduced TG2 activity (Fig. 4E) as well as ϩ We then investigated the mechanisms underlying TG2 up-regula- p42/44 phosphorylation (Fig. 4E) in IB3-1 cells. The Ca2 tion in CFTR-defective cells. Because TG2 activity is strictly Ca2ϩ ionophore ionomycin (27), on the contrary, increased TG2 ac- dependent (22), and high Ca2ϩ mobilization has been shown in CF tivity (Fig. 4E) and phospho-p42/p44 (Fig. 4E) in C38 cells. nasal or bronchial epithelia (26), we looked at the role of Ca2ϩ in These results demonstrated that Ca2ϩ ions drive TG2 activation CFTR-defective cells. and thus inflammation in CF epithelium. 7702 TG2 DRIVES INFLAMMATION VIA PPAR␥ DOWN-REGULATION

FIGURE 4. Inhibition of TG2 activity controls inﬂammation in human airway CF epithelia and CFTR-defective cells. A, Confocal images of phospho-p42/p44 MAPK expression (green) by CF nasal mucosa and IB3-1 cells cultured with or without R283; CyTRAK Orange (red) nuclear counterstaining. B, Immunoblot of phospho-p42/p44 MAPKs in IB3-1 cells after TG2 gene silencing or incubation with R283. Scramble siRNA was used as a negative control. C, TNF-␣ ELISA in IB3-1 cells after TG2 silencing or incubation with R283. The experiment was repeated p Ͻ 0.05 vs IB3-1 cells ,ء) three times cultured in medium alone). Error bars represent SEMs. Confocal images of D, phospho-tyr expression (Py99 Ab) by CF nasal mucosa after incubation with R283 or R283, followed by GW9662, and E, TG2 activity (blue) in IB3-1 cells with Ca2ϩ-chelator BAPTA-AM and in C38 cells with Ca2ϩ-ionophore ionomycin. Phospho- p42/p44 MAPK expression (green) by IB3-1 cells cultured with or without BAPTA-AM and by C38 cells cultured with or without ionomycin. CyTRAK Orange (red) nuclear counterstaining. F and G, NAC decreases expression of TG2 activity (blue), phospho-tyr, phospho-p42/p44 (green) in CF nasal polyp mucosa (F), and in IB3-1 cells (G). Expression of TG2 protein (green) in both CF nasal polyp mucosa and IB3-1 is not affected by NAC treatment (F and G). H, TNF-␣ secretion in IB3-1 cell line cultured for 48 h with or without NAC. The experiment was repeated p Ͻ 0.05 vs IB3-1 cells ,ء) three times cultured in medium alone). Error bars represent SEMs. A and D–G, Scale bar, 10 ␮m.

Because TG2 activity is also regulated by ROS (27), we In agreement with previous reports, high levels of ROS were tested whether ROS influenced the TG2-induced inflammation only detected in CF tissues and CFTR-defective cells (data not in CF. shown) (28). The ROS scavenger NAC significantly reduced TG2 The Journal of Immunology 7703

FIGURE 5. CTR inhibition induces airways inﬂam- mation. The 16HBE cells were cultured with or without CFTRinh-172 inhibitor. A, Confocal microscopy of intracellular ROS (CM-H2DCFDA detection, green) or a Wallac 1420 multilabel counter (each bar represents the mean plus SEMs of three separate experiments, each ,p Ͻ 0.01 compared with C38 cells). B ,ء ;with n ϭ 8 Confocal images of TG2 (green) and its activity (blue). Merge indicates overlay of individual channels (cyan). C, Confocal images of PPAR␥ and phospho-p42/44 MAPKs (green); CyTRAK Orange (red) nuclear counterstaining. Immunoblot analysis of PPAR␥ (55 kDa) and phospho-p42/44 MAPKs (42 and 44 kDa). A–C, Scale bar, 10 ␮m.

activity (Fig. 4, F and G), phosphotyrosine expression (Fig. 4F), Discussion G ␣ H p42/44 phosphorylation (Fig. 4 ), and TNF- release (Fig. 4 )in The novel finding of the present study is the demonstration both CF tissues and IB3-1 cells. Similar results were obtained after that CFTR mutations lead to an increase of TG2 levels and inhibition of ROS by another ROS scavenger, EUK 134 (data not activity that, in turn, down-regulate the anti-inflammatory ef- shown). fects of PPAR␥.IncftrϪ/Ϫ mice, PPAR␥ expression is down- CFTR inhibition induces TG2 up-regulation and regulated both at the mRNA and protein levels, and its function inflammation in normal bronchial 16HBE cells is reduced compared with wild-type littermates (29). Therefore, PPAR␥ agonists have been exploited in therapeutic approaches To assess that the pathway described to date was a direct consequence of CFTR dysfunction, we blocked the functional CFTR in to control inflammation in airway diseases in animal models the normal bronchial 16HBE cell line with the selective inhibitor (30, 31). ␥ CFTRinh-172 (8). We observed all the modifications we have We have shown that PPAR is not only reduced in CF epi- reported above, such as increase of intracellular ROS (Fig. 5A), thelium and CFTR-defective cell lines, but also confined to pe- increase of TG2 protein and its activity (Fig. 5B), decrease of rinuclear aggresomes, prominent pathological features common PPAR␥ (Fig. 5C), and increase of p42/44 phosphorylation to many neurodegenerative conditions such as Huntington and (Fig. 5C). Parkinson’s diseases (32). Aggresomes are generated in re- Altogether, these results clearly demonstrated how CFTR dis- sponse to alterations of the ubiquitin proteasome system, the ruption directly leads to the development of airway inflammation principal cellular mechanism to eliminate misfolded proteins in CF. before they aggregate (33). Aggresomes often develop when a 7704 TG2 DRIVES INFLAMMATION VIA PPAR␥ DOWN-REGULATION threshold of misfolded protein is exceeded, and are usually en- 2. Smith, J. J., S. M. Travis, E. P. Greenberg, and M. J. Welsh. 1996. Cystic fibrosis riched of molecular chaperone, proteasome subunits, ubiquitin airway epithelia fail to kill bacteria because of abnormal airway surface fluid. Cell 85: 229–236. conjugates, and the histone deacetylase HDAC6. In CF, the 3. Kowalski, M. P., A. Dubouix-Bourandy, M. Bajmoczi, D. E. Golan, T. Zaidi, defective CFTR is often not transported beyond the endoplas- Y. S. Coutinho-Sledge, M. P. Gygi, S. P. Gygi, E. A. Wiemer, and G. B. Pier. 2007. Host resistance to lung infection mediated by major vault protein in epi- mic reticulum and accumulates in aggresomes (16, 33). The data thelial cells. Science 6: 130–132. shown in Fig. 1C clearly indicate that also PPAR␥ colocalizes in 4. Thelin, W. R., and R. C. Boucher. 2007. The epithelium as a target for therapy aggresomes with ubiquitin and HDAC6. In CFTR-defective cells, the in cystic fibrosis. Curr. Opin. Pharmacol. 7: 290–295. 5. Osika, E., J. M. Cavaillon, K. Chadelat, M. Boule, C. Fitting, G. Tournier, and specific inhibition of the ubiquitin-proteasome system with MG-132 A. Clement. 1999. Distinct sputum cytokine profiles in cystic fibrosis and other led to a significant increase of PPAR␥, HDAC6-, and ubiquitin- chronic inflammatory airway disease. Eur. Respir. J. 14: 339–346. PPAR␥ colocalization in aggresomes (Fig. 1, D–F). This suggests that 6. Dubin, P. J., F. McAllister, and J. K. Kolls. 2007. Is cystic fibrosis a TH17 disease? Inflamm. Res. 56: 221–227. the formation of these aggresomes in CF is caused by the accumula- 7. Verhaeghe, C., C. Remouchamps, B. Hennuy, A. Vanderplasschen, A. Chariot, tion of misfolded proteins fated to final degradation by the protea- S. P. Tabruyn, C. Oury, and V. Bours. 2007. Role of IKK and ERK pathways in some-ubiquitination system. intrinsic inflammation of cystic fibrosis airways. Biochem. Pharmacol. 73: 1982–1994. The report that in Parkinson’s disease TG2 cross-links 8. Perez, A., A. C. Issler, C. U. Cotton, T. J. Kelley, A. S. Verkman, and P. B. Davis. ␣-synuclein and induces its aggregation in high molecular mass 2007. CFTR inhibition mimics the cystic fibrosis inflammatory profile. aggresomes (34) prompted us to investigate whether TG2 also Am. J. Physiol. 292: L383–L395. 9. Raia, V., L. Maiuri, C. Ciacci, I. Ricciardelli, L. Vacca, S. Auricchio, had a role in the aggresome formation in CF. TG2 is a pleio- M. Cimmino, M. Cavaliere, M. Nardone, A. Cesaro, et al. 2005. Inhibition of p38 tropic enzyme with a calcium-dependent transamidating activ- mitogen activated protein kinase controls airway inflammation in cystic fibrosis. ity that results in cross-linking of proteins via ␧(␥-glutamyl) Thorax 60: 773–780. 10. Egan, M. E., J. Glo¨ckner-Pagel, C. Ambrose, P. A. Cahill, L. Pappoe, lysine bonds (22, 23). Our results clearly demonstrated a sig- N. Balamuth, E. Cho, S. Canny, C. A. Wagner, J. Geibel, and M. J. Caplan. nificant increase of TG2 protein and enzymatic activity in CF 2002. Calcium-pump inhibitors induce functional surface expression of Delta epithelium and CFTR-defective cell lines. TG2 up-regulation F508-CFTR protein in cystic fibrosis epithelial cells. Nat. Med. 8: 485–492. 11. Daynes, R. A., and D. C. Jones. 2002. Emerging roles of PPARs in inflammation was also responsible for the aggresome formation in CF. Im- and immunity. Nat. Rev. Immunol. 2: 748–759. portantly, TG2 induced cross-linking of PPAR␥ and its seques- 12. Maiuri, L., C. Ciacci, I. Ricciardelli, L. Vacca, V. Raia, A. Rispo, M. Griffin, tration into aggresomes (Fig. 3, A–D), because the TG2-inhib- T. Issekutz, S. Quaratino, and M. Londei. 2005. Unexpected role of surface transglutaminase type II in celiac disease. Gastroenterology 129: 1400–1413. itor R283 produced a significant reduction of PPAR␥ protein 13. Bailey, S. T., and S. Ghosh. 2005. ‘PPAR’ting ways with inflammation. Nat. aggregates and restored a wider intracellular protein distribu- Immunol. 6: 966–967. tion (Fig. 3, E–G). We have proven the unequivocal role of TG2 14. Kelly, D., J. I. Campbell, T. P. King, G. Grant, E. A. Jansson, A. G. Coutts, S. Pettersson, and S. Conway. 2004. Commensal anaerobic gut bacteria attenuate in controlling PPAR␥ and inflammation by blocking TG2 en- inflammation by regulating nuclear cytoplasmic shuttling of PPAR-␥ and RelA. zymatic activity (R283) as well as its translation (siRNA). Fur- Nat. Immunol. 5: 104–112. 15. Collino, M., M. Aragno, R. Mastrocola, M. Gallicchio, A. C. Rosa, thermore, the inhibition of the normal CFTR in a normal C. Dianzani, O. Danni, C. Thiemermann, and R. Fantozzi. 2006. Modulation 16HBE bronchial cell line led to a dramatic up-regulation of of the oxidative stress and inflammatory response by PPAR-␥ agonists in the TG2 and inflammation, clearly showing the direct link be- hippocampus of rats exposed to cerebral ischemia/reperfusion. Eur. J. Phar- macol. 530: 70–80. tween CFTR defectiveness and TG2 up-regulation and 16. Kopito, R. R. 2000. Aggresomes, inclusion bodies and protein aggregation. inflammation. Trends Cell Biol. 10: 524–530. Collectively, our study provides a molecular explanation of the 17. Kawaguchi, Y., J. J. Kovacs, A. McLaurin, J. M. Vance, A. Ito, and T. P. Yao. 2003. The deacetylase HDAC6 regulates aggresome formation and cell viability reported association between CFTR malfunction and sterile in- in response to misfolded protein stress. Cell 115: 727–738. flammation. Highlighting the role of TG2 in CF, we have also 18. Hauser, S., G. Adelmant, P. Sarraf, H. M. Wright, E. Mueller, and indicated novel ways to modulate inflammation, a key component B. M. Spiegelman. 2000. Degradation of the peroxisome proliferator-activated receptor ␥ is linked to ligand dependent activation. J. Biol. Chem. 275: in CF pathogenesis. 18527–18533. Furthermore, increased amounts of TG2 have been recently 19. Taylor, J. P., J. Hardy, and K. H. Fischebeck. 2002. Toxic proteins in neurode- described in other diseases, including cancer, in which up-reg- generative disease. Science 296: 2354–2360. 20. Zainelli, G. M., C. A. Ross, J. C. Troncoso, and N. A. Muma. 2003. Transglu- ulation of TG2 is associated with an increased metastatic ac- taminase cross-links in intranuclear inclusions in Huntington disease. J. Neuro- tivity (35) or drug resistance, as in breast cancer, through the pathol. Exp. Neurol. 62: 14–24. activation of NF-␬B (36, 37). Because of its capacity to activate 21. Andringa, G., K. Y. Lam, M. Chegary, X. Wang, T. N. Chase, and M. C. Bennett. 2004. Tissue transglutaminase catalyzes the formation of ␣-synuclein cross-links NF-␬B, a crucial mediator of inflammation-induced tumor in Parkinson’s disease. FASEB J. 18: 932–934. growth and metastatic progression, through depletion of free 22. Lorand, L., and R. M. Graham. 2003. Transglutaminases: cross-linking enzymes I␬B␣ (38), TG2 provides a mechanistic link between inflam- with pleiotropic functions. Nat. Rev. Mol. Cell Biol. 4: 140–156. 23. Siegel, M., and C. Khosla. 2007. Transglutaminase 2 inhibitors and their thera- mation and cancer. peutic role in disease states. Pharmacol. Ther. 115: 232–245. In this view, our study may have a wider impact in the whole 24. Skill, N. J., T. S. Johnson, I. G. Coutts, R. E. Saint, M. Fisher, L. Huang, area of inflammation and cancer. A. M. El Nahas, R. J. Collighan, and M. Griffin. 2004. Inhibition of transglutaminase activity reduces extracellular matrix accumulation induced by high glucose levels in proximal tubular epithelial cells. J. Biol. Chem. 279: 47754–47762. Acknowledgments 25. Belvisi, M. G., D. J. Hele, and M. A. Berrel. 2006. Peroxisome proliferator- activated receptor ␥ agonists as therapy for chronic airway inflammation. Eur. We thank Martin Griffin (Aston University, Birmingham, U.K.) and Chai- J. Pharmacol. 533: 101–109. tan Khosla (Stanford University, Palo Alto, CA) for the gift of the TG2- 26. Ribeiro, C. M., A. M. Paradiso, M. A. Carew, S. B. Shears, and R. C. Boucher. specific inhibitors R283 and KCC009, respectively. 2005. Cystic fibrosis airway epithelial Ca2ϩ i signaling: the mechanism for the larger agonist-mediated Ca2ϩ signals in human cystic fibrosis airway epithelia. J. Biol. Chem. 280: 10202–10209. 27. Yi, S. J., K. H. Kim, H. J. Choi, J. O. Yoo, H. I. Jung, J. A. Han, Y. M. Kim, Disclosures I. B. Suh, and K. S. Ha. 2006. [Ca(2ϩ)]-dependent generation of intracellular The authors have no financial conflict of interest. reactive oxygen species mediates maitotoxin-induced cellular responses in human umbilical vein endothelial cells. Mol. Cells 21: 121–128. 28. Starosta, V., E. Rietschel, K. Paul, U. Baumann, and M. Griese. 2006. Oxidative changes of bronchoalveolar proteins in cystic fibrosis. Chest 129: 431–437. References 29. Ollero, M., O. Junaidi, M. M. Zaman, I. Tzameli, A. A. Ferrando, C. Andersson, 1. Ratjen, F., and G. Doring. 2003. Cystic fibrosis. Lancet 361: 681–689. P. G. Blanco, E. Bialecki, and S. D. Freedman. 2004. Decreased expression of The Journal of Immunology 7705

peroxisome proliferator activated receptor ␥ in cftrϪ/Ϫ mice. J. Cell. Physiol. for Lewy body formation in Parkinson’s disease and dementia with Lewy bodies. 200: 235–244. Proc. Natl. Acad. Sci. USA 100: 2047–2052. 30. Rizzo, G., and S. Fiorucci. 2006. PPARs and other nuclear receptors in inflam- 35. Satpathy, M., L. Cao, R. Pincheira, R. Emerson, R. Bigsby, H. Nakshatri, and mation. Curr. Opin. Pharmacol. 6: 421–427. D. Matei. 2007. Enhanced peritoneal ovarian tumor dissemination by tissue trans- 31. Lee, K. S., S. J. Park, S. R. Kim, K. H. Min, S. M. Jin, H. K. Lee, and Y. C. Lee. glutaminase. Cancer Res. 67: 7194–7202. 2006. Modulation of airway remodeling and airway inflammation by peroxisome 36. Antonyak, M. A., A. M. Miller, J. M. Jansen, J. E. Boehm, C. E. Balkman, proliferator-activated receptor ␥ in a murine model of toluene diisocyanate-in- J. J. Wakshlag, R. L. Page, and R. A. Cerione. 2004. Augmentation of tissue duced asthma. J. Immunol. 177: 5248–5257. transglutaminase expression and activation by epidermal growth factor inhibit 32. Taylor, J. P., F. Tanaka, J. Robitschek, C. M. Sandoval, A. Taye, doxorubicin-induced apoptosis in human breast cancer cells. J. Biol. Chem. 279: S. Markovic-Plese, and K. H. Fischbeck. 2003. Aggresomes protect cells by 41461–41467. enhancing the degradation of toxic polyglutamine containing protein. Hum. Mol. 37. Kim, D. S., S. S. Park, B. H. Nam, I. H. Kim, and S. Y. Kim. 2006. Reversal of Genet. 12: 749–757. drug resistance in breast cancer cells by transglutaminase 2 inhibition and nuclear 33. Johnston, J. A., C. L. Ward, and R. R. Kopito. 1998. Aggresomes: a cellular factor-␬B inactivation. Cancer Res. 66: 10936–10943. response to misfolded proteins. J. Cell Biol. 143: 1883–1898. 38. Karin, M., and F. R. Greten. 2005. NF-␬B: linking inflammation and 34. Juun, E., R. D. Ronchetti, M. M. Quezado, S. Y. Kim, and M. M. Mouradian. immunity to cancer development and progression. Nat. Rev. Immunol. 5: 2003. Tissue transglutaminase induced aggregation of ␣ synuclein: implication 749–759.